Introduction

We thank you for the opportunity to comment on this important notice of inquiry (NOI)[1] on artificial intelligence (AI) and copyright. We appreciate the U.S. Copyright Office undertaking a comprehensive review of the policy and copyright-law issues raised by recent advances in generative AI systems. This NOI covers key areas that require attention, from legal questions regarding infringement and fair use, to questions about how policy choices could shape opportunities for creators and AI producers to engage in licensing.

At this early date, AI systems have already generated some incredible visual art and impressive written texts, as well as a good deal of controversy. Some artists have banded together as part of an anti-AI campaign;^[2] lawsuits have been filed;^[3] and policy experts have attempted to think through the various legal questions raised by these machine-learning systems.

The debates over the role of AI in creative industries have particular salience for intellectual-property rights. Copyright is notoriously difficult to protect online, and the emergence of AI may exacerbate that difficulty. AI systems also potentially pose an additional wrinkle: it is at least arguable that the outputs they produce can themselves be considered unique creations. There are, of course, other open questions whose answers are relevant here, not the least being whether it is fair to assert that only a human can be “creative” (at least, so far).[4]

But leaving these questions aside, we can say that at least some AI systems produce unique outputs and are not merely routinely duplicating other pieces of work in a digital equivalent of collage. That is, at some level, the machines are engaged in a rudimentary sort of “learning” about how humans arrange creative inputs when generating images, music, or written works. The machines appear to be able to reconstruct this process and produce new sets of words, sounds, or lines and colors that conform to the patterns found in human art, in at least a simulacrum of “creativity.”

But that conclusion isn’t the end of the story. Even if some of these AI outputs are unique and noninfringing, the way that AI systems learn—by ingesting massive quantities of existing creative work—raises a number of thorny copyright-law issues. Indeed, some argue that these systems inherently infringe copyright during the learning phase and that, as discussed below, such processes may not survive a “fair use” analysis.

But nor is that assertion the end of the analysis. Rather, it raises the question of whether applying existing doctrine in this novel technological context yields the best results for society. Moreover, it heightens the need for a comprehensive analytical framework to help parse these questions.

A.            The Law & Economics of Copyright and AI

Nearly all would agree that it is crucial that law and public policy strike the appropriate balance between protecting creators’ existing rights and enabling society to enjoy the potentially significant benefits that could arise from the development of AI systems. Indeed, the subject is often cast as a dramatic conflict between creative professionals struggling to make ends meet and innovative firms working to provide cutting-edge AI technology. For the moment, however, it is likely more important to determine the right questions to ask and the proper analytical framework to employ than it is to identify any precise balancing point.

What is important to remember is that copyright policy is foremost economic in nature and “can be explained as a means for promoting efficient allocation of resources.”[5] That is to say, the reason that property rights in creative expression exist is to guarantee the continued production of such works.[6] The fundamental tradeoff in copyright policy is between the costs of limiting access to creative works, and the value obtained by encouraging production of such works.[7] The same applies in the context of AI: identifying the key tradeoffs and weighing the costs and benefits of restricting access to protected works by the producers (and users) of AI systems.[8]

This entails examining the costs and benefits of relatively stronger or weaker forms copyright protection in terms of their effects on both incentives and access, and as they relate to both copyright holders and AI-system developers. It also requires considering where the transaction costs should be allocated for negotiating access to both copyright and, as discussed infra,[9] the use of name/image/likeness, as well as how those allocations are likely to shape outcomes.

At root, these questions center on how to think about the property rights that limit access to protected works and, possibly even more importantly, how to assign new property rights governing the ability to control the use of a name/image/likeness. As we know from the work of the late Nobel laureate Ronald Coase, the actual demarcation of rights affects parties’ abilities to negotiate superior solutions.^[10] The development of nuisance law provides a good example of the problem at hand. When a legal regime provides either strict liability or no-liability rules around pollution, parties have little incentive to minimize harmful conduct:

The factory that has the absolute right to pollute will, if transaction costs are prohibitive, have no incentives to stop (or reduce) pollution even if the cost of stopping would be much less than the cost of pollution to the homeowners. Conversely, homeowners who have an absolute right to be free from pollution will, if transaction costs are prohibitive, have no incentive to take steps of their own to reduce the effects of pollution even if the cost to them of doing so (perhaps by moving away) is less than the cost to the factory of not polluting or of polluting less.^[11]

As Coase observed, this class of problem is best regarded as reciprocal in nature, and the allocation of rights matters in obtaining an efficient outcome. This is necessarily so because, when fully considered, B’s ability to restrain A from the pollution-generating activity can itself be conceived of as another kind of harm that B can impose on A. Therefore, the problem requires a balancing of the relative harms generated by both A and B in exercising conflicting claims in a particular context.

When thinking about how to minimize harms—whether from pollution or other activity that generates social costs (which is to say, nearly every activity)—the aim is to decide whether “the gain from preventing the harm is greater than the loss which would be suffered elsewhere as a result of stopping the action which produces the harm.”^[12] Theoretically, in a world without transaction costs, even assignments of no-liability or strict-liability rules could be bargained around. But we do not live in such a world.^[13] Thus, “[i]n a world in which there are costs of rearranging the rights established by the legal system [common law and statutory assignments of liability] are, in effect, making a decision on the economic problem and determining how resources are to be employed.”^[14]

While pollution rules, unlicensed uses of intellectual property, and a host of other activities subject to legal sanction are not typically framed as resource-allocation decisions, it is undeniable that they do have this character. This is true even where legislation attempts to correct deficiencies in the system. We experience a form of blindness when we focus on correcting what may be rightly perceived as problems in a liability regime. Such analysis tends to concentrate attention on particular deficiencies of the system and to nourish the belief that any measure that removes the deficiency is necessarily desirable. It diverts attention from other changes inevitably associated with the corrective measure—changes that may well produce more harm than the original deficiency.^[15]

All of this is to say that one solution to the costs generated by the need for AI systems to process a massive corpus of expensive, copyright-protected material is neither to undermine property rights, nor to make AI impossible, but to think about how new property rights could make the system work. It may be that some entirely different form or allocation of property right would facilitate bargaining between rightsholders and AI creators, optimizing resource allocation in a way the existing doctrinal regime may not be able to.

A number of other questions flow from this insight into the allocative nature of copyright. How would the incentives for human creators change under different copyright rules for AI systems, or in the face of additional rights? And how would access to copyrighted works for AI training change with different rules, and what effects would that access have on AI innovation?

Above all, our goal today should be to properly frame the AI and copyright debate by identifying tradeoffs, quantifying effects (where possible), and asking what rules best serve the overall objectives of the copyright system and the social goal of encouraging AI innovation. The best chance of striking the right balance will come from a rigorous framing of the questions and from the use of economic analysis to try to answer them.

B.            Copyright Law and AI: Moving Forward

As the Copyright Office undertakes this inquiry, it is important to recognize that, regardless of how the immediate legal questions around AI and copyright are resolved, the growing capabilities and adoption of generative AI systems will likely necessitate some changes in the long term.

The complex questions surrounding the intersection of AI and copyright law admit reasonable arguments on both sides. But AI is here to stay, regardless, and if copyright law is applied in an unduly restrictive manner that substantially hinders socially beneficial AI innovation, it could provoke a broader public-policy backlash that does more to harm copyright’s ability to protect creative works than it does to stanch AI’s ability to undermine it. Copyright law risks being perceived as an obstruction to technological progress if it is used preemptively to kill AI in the cradle. Such an outcome could galvanize calls for recalibrating copyright’s scope and protections in the name of the public interest.

This illustrates the precarious balancing act that copyright law faces in the wake of rapidly evolving technologies like AI. Aggressive copyright restrictions that curtail AI development could instigate a public-policy counter-reaction before Congress and the courts that ultimately undermines copyright’s objectives. The judicious course is to adapt copyright law cautiously to enable AI’s responsible evolution, while resolutely preserving the incentives for human creativity.

In the remainder of this analysis, we offer our perspective on the likely outcomes of the AI-copyright issues raised in this NOI, given the current state of the law. These assessments reflect our perspective formed through the rigorous application of established copyright principles and precedent to the novel technological context of generative AI systems. Reasonable arguments rooted in existing doctrine could be made to support different conclusions. We submit these comments not as definitive predictions or normative preferences, but rather as informed appraisals of how courts may analyze AI under present copyright law, absent legislative intervention.

We appreciate the Copyright Office starting this process to modernize copyright law for the AI age. This inquiry is an important first step, but openness to further evolution will be key to promoting progress in both AI and the arts. We believe an open, evidence-based discussion of these issues will lead to balanced solutions that uphold copyright’s constitutionally mandated purpose, while allowing responsible AI innovation for the public benefit.

II.            The Training of AI Systems and the Applicability of Fair Use

In the NOI, the Copyright Offices asks: “[u]nder what circumstances would the unauthorized use of copyrighted works to train AI models constitute fair use?”^[16]

To answer this question, it would be useful to first briefly walk through a high-level example of how AI systems work, in order to address the most relevant points of contact between AI systems and copyright law.

A.            A Brief Technical Description of AI Training

AI-generated content is not a single “thing,” but a collection of differing processes, each with different implications for the law. For the purposes of this discussion, we will discuss image generation using “generated adversarial networks” (GANs) and diffusion models. Although different systems and different types of content generation will vary, the basic concepts discussed below are nonetheless useful at a general level.^[17]

A GAN is a type of machine-learning model that consists of two parts: a generator and a discriminator.^[18] The generator is trained to create new images that look like they come from a particular dataset, while the discriminator is trained to distinguish the generated images from real images in its original dataset.^[19] The two parts are trained together in an adversarial manner, with the generator trying to produce images that can fool the discriminator and the discriminator trying to correctly identify the generated images.^[20]

A diffusion model, by contrast, analyzes the distribution of information in an image, as noise is progressively added to it.^[21] This kind of algorithm analyzes characteristics of sample images, like the distribution of colors or lines, in order to understand what counts as an accurate representation of a subject (i.e., what makes a picture of a cat look like a cat, and not like a dog).^[22]

For example, in the generation phase, diffusion-based systems start with randomly generated noise, and work backward in “denoising” steps to essentially “see” shapes:

The sampled noise is predicted so that if we subtract it from the image, we get an image that’s closer to the images the model was trained on (not the exact images themselves, but the distribution – the world of pixel arrangements where the sky is usually blue and above the ground, people have two eyes, cats look a certain way – pointy ears and clearly unimpressed).^[23]

While it is possible that some implementations might be designed in a way that saves copies of the training images,^[24] for at least some systems, once the network is trained using these techniques, it will not need to rely on saved copies of input work in order to produce outputs. The models that are produced during training are, in essence, instructions to a different piece of software about how to start with a prompt from a user, a palette of pure noise, and progressively “discover” signal in that image until some new image emerges.

B.            Fair Use

The creator of some of the most popular AI tools, OpenAI, is not shy about their use of protected works in the training phase of the algorithms. In comments to the U.S. Patent and Trademark Office (PTO), OpenAI noted that:

Modern AI systems require large amounts of data. For certain tasks, that data is derived from existing publicly accessible “corpora”… of data that include copyrighted works. By analyzing large corpora (which necessarily involves first making copies of the data to be analyzed), AI systems can learn patterns inherent in human-generated data and then use those patterns to synthesize similar data which yield increasingly compelling novel media in modalities as diverse as text, image, and audio. (emphasis added).^[25]

Thus, at the training stage, the most popular forms of AI systems require making copies of existing works. And where that material is either not in the public domain or is not licensed, an infringement can occur. Thus, the copy must not be infringing (say, because it is transient), or some affirmative defense is needed to excuse the infringement. Toward this end, OpenAI believes that this use should qualify as fair use,^[26] as do most or all the other major producers of generative AI systems.^[27]

But as OpenAI has framed the fair-use analysis, it is not clear that these uses should qualify. There are two major questions in this respect: will the data used to train these systems count as “copies” under the Copyright Act, and, if so, is the use of these “copies” sufficiently “transformative” to qualify for the fair-use defense?

1.              Are AI systems being trained with ‘copies’ of protected works?

Section 106 of the Copyright Act grants the owner of a copyright the exclusive right “to reproduce… copyrighted work in copies” and to authorize others to do so.^[28] If an AI system makes a copy of a file to a computer during training, this would likely constitute a prima facie violation of the copyright owner’s exclusive right of reproduction under Section 106. This is fairly straightforward.

But what if the “copy” is “transient” and/or only partial pieces of content are used in the training? For example, what if a training program merely streamed small bits of a protected work into temporary memory as part of its training, and retained no permanent copy?

As the Copyright Office has previously observed, even temporary reproductions of a work in a computer’s memory can constitute “copies” under the Copyright Act.^[29] Critically, this includes even temporary reproductions made as part of a packet-switching network transmission, where a particular file is broken into individual packets, because the packets can be reassembled into substantial portions or even entire works.^[30] On the topic of network-based transmission, the Copyright Office further observed that:

Digital networks permit a single disk copy of a work to meet the demands of many users by creating multiple RAM copies. These copies need exist only long enough to be perceived (e.g., displayed on the screen or played through speakers), reproduced or otherwise communicated (e.g., to a computer’s processing unit) in order for their economic value to be realized. If the network is sufficiently reliable, users have no need to retain copies of the material. Commercial exploitation in a network environment can be said to be based on selling a right to perceive temporary reproductions of works.^[31]

This is a critical insight that translates well to the context of AI training. The “transience” of the copy matters with respect to the receiver’s ability to perceive the work in a way that yields commercial value. Under this reasoning, the relevant locus of analysis is on the AI system’s ability to “perceive” a work for the purposes of being trained to “understand” the work. In this sense, you could theoretically find the existence of even more temporary copies than that necessary for human perception to implicate the reproduction right.

Even where courts have been skeptical of extending the definition of “copy” to “fleeting” copies in computer memory, this underlying logic is revealed. In Cartoon Network LP, LLLP v. CSC Holdings, Inc., 536 F.3d 121 (2008), the 2nd U.S. Circuit Court of Appeals had to determine whether buffered media sent to a DVR device was too “transient” to count as a “copy”:

No bit of data remains in any buffer for more than a fleeting 1.2 seconds. And unlike the data in cases like MAI Systems, which remained embodied in the computer’s RAM memory until the user turned the computer off, each bit of data here is rapidly and automatically overwritten as soon as it is processed. While our inquiry is necessarily fact-specific, and other factors not present here may alter the duration analysis significantly, these facts strongly suggest that the works in this case are embodied in the buffer for only a “transitory” period, thus failing the duration requirement.^[32]

In Cartoon Network, the court acknowledged both that the duration analysis was fact-bound, and also that the “fleeting” nature of the reproduction was important. “Fleeting” is a relative term, based on the receiver’s capacities. A ball flying through the air may look “fleeting” to a human observer, but may appear to go much more cognizable to a creature with faster reaction time, such as a house fly. So, too, with copies of a work in a computer’s memory and the ability to “perceive” what is fixed in a buffer: what may be much too quick for a human to perceive may very well be within an AI system’s perceptual capabilities.

Therefore, however the training copies are held, there is a strong possibility that a court will find them to be “copies” for the purposes of the reproduction right—even with respect to partial copies that exist for very small amounts of time.

2.              The purpose and character of using protected works to train AI systems

Fair use provides for an affirmative defense against infringement when the use is, among other things, “for purposes such as criticism, comment, news reporting, teaching…, scholarship, or research.”^[33] When deciding whether a fair-use defense is applicable, a court must balance a number of factors:

1. the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
2. the nature of the copyrighted work;
3. the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
4. the effect of the use upon the potential market for or value of the copyrighted work.^[34]

The fair-use defense that AI creators have advanced is rooted in the first factor: the nature and character of the use. Although a full analysis of all the factors is ultimately necessary, analysis of the first factor is sufficiently complicated to warrant full attention here. In particular, the complex issue at hand is whether uses of protected works to train AI systems are sufficiently “transformative” or not.^[35]

Whether the use of a copyrighted work to train an AI is “transformative” is certainly a novel question, but it is one that will likely be answered in light of an observation the U.S. Supreme Court made in Campbell v. Acuff Rose Music:

[W]hen a commercial use amounts to mere duplication of the entirety of an original, it clearly “supersede[s] the objects,”… of the original and serves as a market replacement for it, making it likely that cognizable market harm to the original will occur… But when, on the contrary, the second use is transformative, market substitution is at least less certain, and market harm may not be so readily inferred.^[36]

Moreover, “[t]he word ‘transformative’ cannot be taken too literally as a sufficient key to understanding the elements of fair use. It is rather a suggestive symbol for a complex thought, and does not mean that any and all changes made to an author’s original text will necessarily support a finding of fair use.”[37] A key question, then, is whether training AI systems on copyrighted works amounts to a mere “duplication of the entirety of an original” or is sufficiently “transformative” to support a fair-use defense. As noted above, OpenAI believes that its use is transformative. According to its comments:

Training of AI systems is clearly highly transformative. Works in training corpora were meant primarily for human consumption for their standalone entertainment value. The “object of the original creation,” in other words, is direct human consumption of the author’s ?expression.? Intermediate copying of works in training AI systems is, by contrast, “non-expressive” the copying helps computer programs learn the patterns inherent in human-generated media. The aim of this process—creation of a useful generative AI system—is quite different than the original object of human consumption. The output is different too: nobody looking to read a specific webpage contained in the corpus used to train an AI system can do so by studying the AI system or its outputs. The new purpose and expression are thus both highly transformative.^[38]

This framing, however, works against OpenAI’s interests. As noted above, and reinforced in the immediately preceding quote, generative AI systems are made of at least two distinct pieces. The first is a piece of software that ingests existing works and creates a file that can serve as instructions to the second piece of software. The second piece of software takes the output of the first and can produce independent results. Thus, there is a clear discontinuity in the process whereby the ultimate work created by the system is disconnected from the creative inputs used to train the software.

Therefore, the protected works are arguably ingested into the first part of the system “for their standalone entertainment value.” That is to say, the goal of copying and showing a protected work to an AI system is for the analog of “direct human consumption of the author’s expression” in order for the system to learn about that expression.

The software is learning what counts as “standalone entertainment value” and therefore the works must be used in those terms. Surely, a computer is not sitting on a couch and surfing for its pleasure. But it is solely for the very “standalone entertainment value” that the first piece of software is being shown copyrighted material. By contrast, parody or “remixing” uses incorporate a work into some secondary expression that directly transforms the input. The way these systems work is to learn what makes a piece entertaining and then to discard that piece altogether. Moreover, this use for the art qua art most certainly interferes with the existing market, insofar as this use is in lieu of reaching a licensing agreement with rightsholders.

A good analogy is art students and art textbooks. Art students view protected works in an art textbook in order to learn how to reproduce the styles contained therein. The students would not be forgiven for pirating the textbooks merely because they intend to go on to make new paintings. They would still be liable for copyright infringement if they used unlicensed protected works as part of their education.

The 2nd U.S. Circuit Court of Appeals dealt with a case that operates similarly to this dynamic. In American Geophysical Union v. Texaco, 60 F.3d 913 (2d Cir. 1994), the 2nd Circuit considered whether Texaco’s photocopying of scientific articles produced by the plaintiffs qualified for a fair-use defense. Texaco employed between 400 and 500 research scientists and, as part of supporting their work, maintained subscriptions to a number of scientific journals.^[39]

It was common practice for Texaco’s scientists to photocopy entire articles and save them in a file.^[40] The plaintiffs sued for copyright infringement.^[41] Texaco asserted that photocopying by its scientists for the purposes of furthering scientific research—that is to train the scientists on the content of the journal articles—should count as a fair use. The argument was, at least in part, that this was sufficiently “transformative,” because the scientists were using that knowledge to invent new products.^[42] The 2nd Circuit disagreed:

The “transformative use” concept is pertinent to a court’s investigation under the first factor because it assesses the value generated by the secondary use and the means by which such value is generated. To the extent that the secondary use involves merely an untransformed duplication, the value generated by the secondary use is little or nothing more than the value that inheres in the original. Rather than making some contribution of new intellectual value and thereby fostering the advancement of the arts and sciences, an untransformed copy is likely to be used simply for the same intrinsic purpose as the original, thereby providing limited justification for a finding of fair use….^[43]

The 2nd Circuit thus observed that copies of the scientific articles were made solely to consume the material itself. AI developers often make an argument analogous to that made by Texaco: that training AI systems surely advances scientific research, and therefore fosters the “advancement of the arts and sciences.” But in American Geophysical Union, the initial copying of copyrighted content, even where it was ultimately used for the “advancement of the arts and sciences,” was not held to be sufficiently “transformative.”^[44] The case thus stands for the proposition that one cannot merely identify a social goal down that would be advanced at some future date in order to permit an exception to copyright protection. As the court put it:

[T]he dominant purpose of the use is a systematic institutional policy of multiplying the available number of copies of pertinent copyrighted articles by circulating the journals among employed scientists for them to make copies, thereby serving the same purpose for which additional subscriptions are normally sold, or… for which photocopying licenses may be obtained.[45]

The use itself must be transformative and different, and copying is not transformative merely because it may be used as an input into a later transformative use. By the same token, therefore, it seems likely that where an AI system ingests (copies) copyrighted works, that use is similarly not transformative, despite its ultimate use as an input in the creation of other original works.

Comparing the American Geophysical Union analysis with the search-engine “snippets” and “thumbnails” cases provides a useful comparison relevant to the AI analysis. In Kelly v. Arriba Soft Corp., 336 F.3d 811 (9^th Cir. 2002), the 9^th U.S. Circuit Court of Appeals ruled that a search engine’s creation of thumbnail images from original copies was a transformative fair use.[46] Arriba’s search-engine crawler made full-sized copies of Kelly’s images and stored them temporarily on Arriba’s server to generate thumbnail versions. After the thumbnails were created, the full-sized originals were deleted. The thumbnails were used to facilitate Arriba’s image-based search engine. In reaching its fair-use conclusion, the 9^th Circuit opined that:

Arriba’s use of Kelly’s images promotes the goals of the Copyright Act and the fair use exception. The thumbnails do not stifle artistic creativity because they are not used for illustrative or artistic purposes and therefore do not supplant the need for the originals.[47]

Further, although “Arriba made exact replications of Kelly’s images, the thumbnails were much smaller, lower-resolution images that served an entirely different function than Kelly’s original images.”[48]

The court found it important that the search engine did not use the protected works for their intended “aesthetic experience,” but rather for the purpose of constructing a search index.[49] Indeed, the entire point of a search engine is not to “supersede” the original, but in many or most cases to provider users an efficient means to find that original online.[50]

The court discussed, but only briefly, the benefit to the public of Arriba’s transformative use,[51] noting that “[Arriba’s thumbnails] benefit the public by enhancing information-gathering techniques on the internet.”[52] Five years later, in Perfect 10 Inc. v. Amazon.com Inc., 487 F.3d 701 (2007), the 9^th Circuit expanded on this question somewhat.[53] There, in holding that the novelty of the use was of crucial importance to the analysis,[54] the court also stressed that the value of that use was a function of its newness:

[A] search engine provides social benefit by incorporating an original work into a new work, namely, an electronic reference tool. Indeed, a search engine may be more transformative than a parody [the use at issue in Campbell] because a search engine provides an entirely new use for the original work, while a parody typically has the same entertainment purpose as the original work.[55]

Indeed, even in light of the commercial nature of Google’s use of copyrighted content in its search engine, its significant public benefit carried the day: “We conclude that the significantly transformative nature of Google’s search engine, particularly in light of its public benefit, outweighs Google’s superseding and commercial uses of the thumbnails in this case.”[56] And, of particular relevance to these questions in the context of AI, the court in Perfect 10 went on to “note the importance of analyzing fair use flexibly in light of new circumstances.”[57]

Ultimately, the Perfect 10 decision tracked Kelly fairly closely on the rest of the “transformativeness” analysis in finding fair use, because “[a]lthough an image may have been created originally to serve an entertainment, aesthetic, or informative function, a search engine transforms the image into a pointer directing a user to a source of information.”[58]

The core throughline in this line of cases is the question of whether a piece of content is being used for its expressive content, weighed against the backdrop of whether the use is for some new (and, thus, presumptively valuable) purpose. In Perfect 10 and Kelly, the transformative use was the creation of a search index.

“Snippets” fair-use cases track a similar line of reasoning. For example, in Authors Guild v. Google Inc., 804 F.3d 202 (2d Cir. 2015), the 2nd Circuit ruled that Google’s use of “snippets” of copyrighted books in its Library Project and Google Books website was a “transformative” fair use.[59] Holding that the “snippet view” of books digitized as part of the Google Books project did not constitute an effectively competing substitute to the original works, the circuit court noted that copying for the purpose of “criticism” or—as in that case—copying for the purpose of “provision of information about” the protected work, “tends most clearly to satisfy Campbell’s notion of the ‘transformative’ purpose.”[60]

Importantly, the court emphasized the importance of the public-benefit aspect of transformative uses: “[T]ransformative uses tend to favor a fair use finding because a transformative use is one that communicates something new and different from the original or expands its utility, thus serving copyright’s overall objective of contributing to public knowledge.”[61]

Underscoring the idea that the “transformativeness” analysis weighs whether a use is merely for expressive content against the novelty/utility of the intended use, the court observed:

Google’s division of the page into tiny snippets is designed to show the searcher just enough context surrounding the searched term to help her evaluate whether the book falls within the scope of her interest (without revealing so much as to threaten the author’s copyright interests). Snippet view thus adds importantly to the highly transformative purpose of identifying books of interest to the searcher.[62]

Thus, the absence of use of the work’s expressive content, coupled with a fairly circumscribed (but highly novel) use was critical to the outcome.

The entwined questions of transformative use and the public benefit it confers are significantly more complicated in the AI context, however. Unlike the incidental copying involved in search-engine indexing or thumbnails, training generative AI systems directly leverages copyrighted works for their expressive value. In the Google Books and Kelly cases, the defendant systems extracted limited portions of works or down-sampled images solely to identify and catalog their location for search purposes. The copies enabled indexing and access, and they expanded public knowledge through a means unrelated to the works’ protected aesthetics.

But in training AI models on copyrighted data, the systems necessarily parse the intrinsic creative expression of those works. The AI engages with the protected aesthetic elements themselves, not just superficial markers (like title, length, location on the internet, etc.), in order to internalize stylistic and compositional principles. This appropriates the heart of the works’ copyright protection for expressive ends, unlike the more tenuous connections in search systems.

The AI is thus “learning” directly from the protected expression in a manner akin to a human student studying an art textbook, or like the scientists learning from the journals in American Geophysical Union. The subsequent AI generations are built from mastery of the copyrighted training materials’ creative expression. Thus, while search-engine copies only incidentally interact with protected expression to enable unrelated innovation, AI training is predicated on excavating the protected expression itself to fuel iterative creation. These meaningfully different purposes have significant fair-use implications.

This functional difference is, as noted, central to the analysis of a use’s “purpose and character.” Indeed, “even making an exact copy of a work may be transformative so long as the copy serves a different function than the original work.”[63] But the benefit to the public from the new use is important, as well, particularly with respect to the possible legislative response that a restrictive interpretation of existing doctrine may engender.

If existing fair-use principles prohibit the copying required for AI, absent costly item-by-item negotiation and licensing, the transaction costs could become prohibitive, thwarting the development of technologies that promise great public value.[64] Copyright law has faced similar dilemmas before, where the transaction costs of obtaining permission for socially beneficial uses could frustrate those uses entirely.[65] In such cases, we have developed mechanisms like compulsory licensing to facilitate the necessary copying, while still attempting to compensate rightsholders. An unduly narrow fair-use finding for AI training could spur calls for similar interventions in service of enabling AI progress.

In other words, regardless of the veracity of the above conclusion that AI’s use of copyrighted works may not, in fact, serve a different function than the original, courts and legislators may be reluctant to allow copyright doctrine to serve as an absolute bar against self-evidently valuable activity like AI development. Our aim should be to interpret or recalibrate copyright law to permit such progress while upholding critical incentives for creators.

C.            Opt-In vs. Opt-Out Use of Protected Works

The question at the heart of the prior discussion—and, indeed, at the heart of the economic analysis of copyright—is whether the transaction costs that accompany requiring express ex ante permission for the use of protected works are so high that they impedes socially beneficial conduct whose value would outweigh the social cost of allowing permissionless and/or uncompensated use.[66] The NOI alludes to this question when it asks: “Should copyright owners have to affirmatively consent (opt in) to the use of their works for training materials, or should they be provided with the means to object (opt out)?”^[67]

This is a complex problem. Given the foregoing thoughts on fair use, it seems quite possible that, at present, the law requires creators of AI systems to seek licenses for protected content, or else must resort to public-domain works for training. Given the volume of copyrighted works that AI developers currently use to train these systems, such requirements may be broadly infeasible.

On one hand, requiring affirmative opt-in consent from copyright holders imposes significant transaction costs on AI-system developers to identify and negotiate licenses for the vast amounts of training data required. This could hamper innovation in socially beneficial AI systems. On the other hand, an opt-out approach shifts more of the transaction-cost burden to copyright holders, who must monitor and object to unwanted uses of their works. This raises concerns about uncompensated use.

Ultimately, the question is where the burden should lie: with AI-system developers to obtain express consent, or with copyright holders to monitor and object to uses? Requiring some form of consent may be necessary to respect copyright interests. Yet an opt-out approach may strike the right balance, by shifting some of the burden back to AI developers while avoiding the infeasibly high transaction costs of mandatory opt-in consent. The optimal approach likely involves nuanced policymaking to balance these competing considerations. Moreover, as we discuss infra, the realistic outcome is most likely going to require rethinking the allocation of property rights in ways that provide for large-scale licensing. Ideally, this could be done through collective negotiation, but perhaps at a de minimis rate, while allowing creators to bargain for remuneration on the basis of other rights, like a right of publicity or other rights attached to the output of AI systems, rather than the inputs.[68]

1.              Creator consent

Relatedly, the Copyright Office asks: “If copyright owners’ consent is required to train generative AI models, how can or should licenses be obtained?”^[69]

Licensing markets exist, and it is entirely possible that major AI developers and large groups of rightsholders can come to mutually beneficial terms that permit a sufficiently large body of protected works to be made available as training data. Something like a licensing agency for creators who choose to make their works available could arise, similar to the services that exist to provide licensed music and footage for video creators.^[70] It is also possible for some to form collective-licensing organizations to negotiate blanket permissions covering many works.

It’s important to remember that our current thinking is constrained by our past experience. All we know today are AI models trained on vast amounts of unlicensed works. It is entirely possible that, if firms were required to seek licenses, unexpected business models would emerge to satisfy both sides of the equation.

For example, an AI firm could develop its own version of YouTube’s ContentID, which would allow creators to control when their work is used in AI training. For some well-known artists, this could be negotiated with an upfront licensing fee. On the user side, any artist who has opted in could then be selected as a “style” for the AI to emulate—triggering a royalty payment to the artist when a user generates an image or song in that style. Creators could also have the option of removing their influence from the system if they so desire.

Undoubtedly, there are other ways to structure the relationship between creators and AI systems  that would facilitate creators’ monetization of the use of their work in AI systems, including legal and commercial structures that create opportunities for both creators and AI firms to succeed.

III.          Generative AI Outputs: Protection of Outputs and Outputs that Infringe

The Copyright Office asks: “Under copyright law, are there circumstances when a human using a generative AI system should be considered the ‘author’ of material produced by the system?”^[71]

Generally speaking, we see no reason why copyright law should be altered to afford protection to purely automatic creations generated by AI systems. That said, when a human makes a nontrivial contribution to generative AI output—such as editing, reframing, or embedding the AI-generated component within a larger work—the resulting work should qualify for copyright protection.

Copyright law centers on the concept of original human authorship.^[72] The U.S. Constitution expressly limits copyright to “authors.”^[73] As of this writing, however, generative AI’s capacities do not rise to the level of true independent authorship. AI systems remain tools that require human direction and judgment.[74] As such, when a person provides the initial prompt or framing, makes choices regarding the iterative development of the AI output, and decides that the result is satisfactory for inclusion in a final work, they are fundamentally engaging in creative decision making that constitutes authorship under copyright law.

As Joshua Gans has observed of recent Copyright Review Board decisions:

Trying to draw some line between AI and humans with the current technology opens up a massive can of worms. There is literally no piece of digital work these days that does not have some AI element to it, and some of these mix and blur the lines in terms of what is creative and what is not. Here are some examples:

A music artist uses AI to denoise a track or to add an instrument or beat to a track or to just get a composition started.

A photographer uses Photoshop or takes pictures with an iPhone that already uses AI to focus the image and to sort a burst of images into one that is appropriate.

A writer uses AI to prompt for some dialogue when stuck at some point or to suggest a frame for writing a story.[75]

Attempting to separate out an “AI portion” from the final work, as the Copyright Review Board proposed, fundamentally misunderstands the integrated nature of the human-AI collaborative process. The AI system cannot function without human input, and its output remains raw material requiring human creativity to incorporate meaningfully into a finished product.

Therefore, when a generative AI system is used as part of a process guided by human creative choices, the final work should be protected by copyright, just as a work created using any other artistic tool or collaborator would be. Attenuating copyrightability due to the use of AI would undermine basic copyright principles and fail to recognize the essentially human nature of the creative process.

A.            AI Outputs and Infringement

The NOI asks: “Is the substantial similarity test adequate to address claims of infringement based on outputs from a generative AI system, or is some other standard appropriate or necessary?” (Question 23)

The outputs of AI systems may or may not violate IP laws, but there is nothing inherent in the processes described above that dictates that they must. As noted, the most common AI systems do not save copies of existing works, but merely “instructions” (more or less) on how to create new work that conforms to patterns found by examining existing work. If we assume that a system isn’t violating copyright at the input stage, it’s entirely possible that it can produce completely new pieces of art that have never before existed and do not violate copyright.

They can, however, be made to violate copyrights. For example, these systems can be instructed to generate art, not just in the style of a particular artist, but art that very closely resembles existing pieces. In this sense, it would be making a copy that theoretically infringes. The fact of an AI’s involvement would not change the analysis: just as with a human-created work, if it is substantially similar to a copyrighted work, it may be found infringing.

There is, however, a common bug in AI systems that leads to outputs that are more likely to violate copyright in this way. Known as “overfitting,” the training leg of these AI systems can be presented with samples that contain too many instances of a particular image.^[76] This leads to a dataset that contains too much information about the specific image, such that—when the AI generates a new image—it is constrained to producing something very close to the original. Similarly, there is evidence that some AI systems are “memorizing” parts of protected books.^[77] This could lead to AI systems repeating copyright-protected written works.

1.              The substantial-similarity test

The substantial-similarity test remains functionally the same when evaluating works generated using AI. To find “substantial similarity,” courts require evidence of copying, as well as an expression that is substantially similar to a protected work.^[78] “It is now an axiom of copyright law that actionable copying can be inferred from the defendant’s access to the copyrighted work and substantial similarity between the copyrighted work and the alleged infringement.”^[79] In many or most cases, it will arguably be the case that AI systems have access to quite a wide array of protected works that are posted online. Thus, there may not be a particularly high hurdle to determine that an AI system actually copied a protected work.

There is, however, one potential problem for the first prong of this analysis. Models produced during a system’s training process do not (usually) contain the original work, but are the “ideas” that the AI systems generated during training. Thus, where the provenance of works contained in a training corpus is difficult to source, it may not be so straightforward to make inferences about whether a model “saw” a particular work. This is because the “ideas” that the AI “learns” from its training corpus are unprotected under U.S. copyright law, as it is permissible to mimic unprotected elements of a copyrighted work (such as ideas).^[80]

Imagine a generative AI system trained on horror fiction. It would be possible for this system to produce a new short story that is similar to one written by Stephen King, but the latent data in the model almost certainly would not violate any copyrights that King holds in his work. The model would contain “ideas” about horror stories, including those learned from an array of authors who were themselves influences on Stephen King, and potentially some of King’s own stories. What the AI system “learns” in this case is the relationship between words and other linguistic particularities that are commonly contained in horror fiction. That is, it has “ideas” about what goes into a horror story, not (theoretically) the text of the horror story itself.

Thus, when demonstrating indirect proof of copying in the case of a Stephen King story, it may pose a difficulty that an AI system has ingested all of H.P. Lovecraft’s work—an author who had a major influence on King. The “ideas” in the model and the output it subsequently produces may, in fact, produce something similar to a Stephen King work, but it may have been constructed largely or entirely on material from Lovecraft and other public-domain horror writers. The problem becomes only more complicated when you realize that this system could also have been trained on public-domain fan fiction written in the style of Stephen King. Thus, for the purposes of the first prong of this analysis, courts may place greater burden on plaintiffs in copyright actions against model producers to demonstrate more than merely that a work was merely available online.

Assuming that plaintiffs are able to satisfy the first prong, once an AI system “expresses” those ideas, that expression could violate copyright law under the second prong of the substantial-similarity test. The second prong inquires whether the final work appropriated the protected original expression.^[81] Any similarities in unprotectable ideas, facts, or common tropes are disregarded.^[82] So, in both traditional and AI contexts, the substantial-similarity test ultimately focuses on the protected components of creative expression, not surface similarity.

The key determination is whether the original work’s protected expression itself has been impermissibly copied, no matter the process that generated the copy. AI is properly viewed as simply another potential tool that could be used in certain acts of copying. It does not require revisiting settled principles of copyright law.

B.            Direct and Secondary Liability

The NOI asks: “If AI-generated material is found to infringe a copyrighted work, who should be directly or secondarily liable—the developer of a generative AI model, the developer of the system incorporating that model, end users of the system, or other parties?”[83]

Applying traditional copyright-infringement frameworks to AI-generated works poses unique challenges in determining direct versus secondary liability. In some cases, the AI system itself may create infringing content without any direct human causation.

1.              Direct liability

If the end user prompts an AI system in a way that intentionally targets copyrighted source material, they may meet the threshold for direct infringement by causing the AI to reproduce protected expression.^[84] Though many AI prompts contain only unprotected ideas, users may sometimes input copyrightable material as the basis for the AI output. For example, a user could upload a copyrighted image and request the AI to make a new drawing based on the sample. In such cases, the user is intentionally targeting copyrighted works and directly “causing” the AI system to reproduce output that is similar. If sufficiently similar, that output could infringe on the protected input. This would be a question of first impression, but it is a plausible reading of available cases.

For example, in CoStar Grp. Inc. v. LoopNet Inc., 373 F.3d 544 (4th Cir. 2004), the 4^th U.S. Circuit Court of Appeals had to consider whether an internet service provider (ISP) could be directly liable when third parties reposted copyrighted material owned by the plaintiff. In determining that merely owning the “machine” through which copies were made or transmitted was not enough to “cause” a direct infringement, the court held that:

[T]o establish direct liability under §§ 501 and 106 of the Act, something more must be shown than mere ownership of a machine used by others to make illegal copies. There must be actual infringing conduct with a nexus sufficiently close and causal to the illegal copying that one could conclude that the machine owner himself trespassed on the exclusive domain of the copyright owner. The Netcom court described this nexus as requiring some aspect of volition or causation… Indeed, counsel for both parties agreed at oral argument that a copy machine owner who makes the machine available to the public to use for copying is not, without more, strictly liable under § 106 for illegal copying by a customer. The ISP in this case is an analogue to the owner of a traditional copying machine whose customers pay a fixed amount per copy and operate the machine themselves to make copies. When a customer duplicates an infringing work, the owner of the copy machine is not considered a direct infringer. Similarly, an ISP who owns an electronic facility that responds automatically to users’ input is not a direct infringer.^[85]

Implied in the 4^th Circuit’s analogy is that, while the owner of a copying machine might not be a direct infringer, a user employing such a machine could be a direct infringer. It’s an imperfect analogy, but a user of an AI system prompting it to create a “substantially similar” reproduction of a protected work could very well be a direct infringer under this framing. Nevertheless, the analogy is inexact, because the user feeds an original into a copying machine in order to make a more-or-less perfect copy of the original, whereas an AI system generates something new but similar. The basic mechanism of using a machine to try to reproduce a protected work, however, remains essentially the same. Whether there is an infringement would be a question of “substantial similarity.”

2.              Secondary liability

As in the case of direct liability, the nature of generative AI makes the secondary-liability determination slightly more complicated, as well. That is, paradoxically, the basis for secondary liability could theoretically arise even where there was no direct infringement.[86]

The first piece of this analysis is relatively easier. If a user is directly liable for infringing a protected work, as noted above, the developer and provider of a generative AI system may face secondary copyright liability. If the AI developer or distributor knows the system can produce infringing outputs, and provides tools or material support that allows users to infringe, it may be liable for contributory infringement.^[87] Critically, merely designing a system that is capable of infringing is not enough to find contributory liability.^[88]

An AI producer or distributor may also have vicarious liability, insofar as it has the right and ability to supervise users’ activity and a direct financial interest in that activity.^[89] AI producers have already demonstrated their ability to control users’ behavior to thwart unwanted uses of the service.^[90] Thus, if there is a direct infringement by a user, a plausible claim for vicarious liability could be made so long as there is sufficient connection between the user’s behavior and the producer’s financial interests.

The question becomes more complicated when a user did not direct the AI system to infringe. When the AI generates infringing content without user direction, it’s not immediately clear who would be liable for the infringement.^[91] Consider the case where, unprompted by either the user or the AI producer, an AI system creates an output that would infringe under the substantial-similarity test. Assuming that the model has not been directed by the producer to “memorize” the works it ingests, the model itself consists of statistical information about the relationship between different kinds of data. The infringer, in a literal sense, is the AI system itself, as it is the creator of the offending output. Technically, this may be a case of vicarious liability, even without an independent human agent causing the direct infringement.

We know that copyright protection can only be granted to humans. As the Copyright Review Board recently found in a case deciding whether AI-generated outputs can be copyrighted:

The Copyright Act protects, and the Office registers, “original works of authorship fixed in any tangible medium of expression.” 17 U.S.C. § 102(a). Courts have interpreted the statutory phrase “works of authorship” to require human creation of the work.^[92]

But can an AI system directly violate copyright? In his Aereo dissent, Justice Clarence Thomas asserted that it was a longstanding feature of copyright law that violation of the performance right required volitional behavior.^[93] But the majority disagreed with him, holding that, by running a fully automated system of antennas intended to allow users to view video at home, the system gave rise to direct copyright liability.^[94] Thus, implied in the majority’s opinion is the idea that direct copyright infringement does not require “volitional” conduct.

It is therefore plausible that a non-sentient, fully automated AI system could infringe copyright, even if, ultimately, there is no way to recover against the nonhuman agent. That does, however, provide an opportunity for claims of vicarious liability against the AI producer or distributor— at least, where the producer has the power to control the AI system’s behavior and that behavior appears to align with the producer’s financial interests.

3.              Protecting the ‘style’ of human creators

The NOI asks: “Are there or should there be protections against an AI system generating outputs that imitate the artistic style of a human creator (such as an AI system producing visual works ‘in the style of’ a specific artist)?”[95]

At the federal level, one candidate for protection against AI imitating some aspects of a creator’s works can currently be found in trademark law. Trademark law, governed by the Lanham Act, protects names, symbols, and other source identifiers that distinguish goods and services in commerce.^[96] Unfortunately, a photograph or likeness, on its own, typically does not qualify for trademark protection, unless it is consistently used on specific goods.^[97] Even where there is a likeness (or similar “mark”) used consistently as part of branding a distinct product, many trademark-infringement claims would be difficult to establish in this context, because trademark law does little to protect many aspects of a creator’s work.

Moreover, the Supreme Court has been wary about creating a sort of “mutant copyright” in cases that invoke the Lanham Act as a means to enforce a sort of “right of attribution,” which would potentially give creators the ability to control the use of their name in broader contexts.^[98] In this context, the Court has held that the relevant parts of the Lanham Act were not designed to “protect originality or creativity,”^[99] but are focused solely on “actions like trademark infringement that deceive consumers and impair a producer’s goodwill.”^[100]

In many ways, there is a parallel here to the trademark cases involving keyword bidding in online ads. At a high level, search engines and other digital-advertising services do not generally infringe trademark when they allow businesses to purchase ads triggered by a user’s search for competitor trademarks (i.e., rivals’ business names).^[101] But in some contexts, this can be infringing—e.g., where the use of trademarked terms in combination with advertising text can mislead consumers about the origin of a good or service.^[102]

Thus, the harm, when it arises, would not be in a user asking an AI system to generate something “in the style of” a known creator, but when that user subsequently seeks to release a new AI-generated work and falsely claims it originated from the creator, or leaves the matter ambiguous and misleading to consumers.

Alternative remedies for creators could be found in the “right of publicity” laws in various states. A state-level right of publicity “is not merely a legal right of the ‘celebrity,’ but is a right inherent to everyone to control the commercial use of identity and persona and recover in court damages and the commercial value of an unpermitted taking.”^[103] Such rights are recognized under state common law and statutes, which vary considerably in scope across jurisdictions—frequently as part of other privacy statutes.^[104] For example, some states only protect an individual’s name, likeness, or voice, while others also cover distinctive appearances, gestures, and mannerisms.^[105] The protections afforded for right-of-publicity claims vary significantly based on the state where the unauthorized use occurs or the individual is domiciled.^[106] This creates challenges for the application of uniform nationwide protection of creators’ interests in the various aspects that such laws protect.

In recent hearings before the U.S. Senate Judiciary Subcommittee on Intellectual Property, several witnesses advocated creating a federal version of the right of publicity.^[107] The Copyright Office has also previously opined that it may be desirable for Congress to enact some form of a “right of publicity” law.^[108] If Congress chose to enact a federal “right of privacy” statute, several key issues would need to be addressed regarding the scope of protection, effect on state laws, constitutional authority, and First Amendment limitations.

Congress would have to delineate the contours of the federal right of publicity, including the aspects of identity covered and the types of uses prohibited. A broad right of privacy could protect names, images, likenesses, voices, gestures, distinctive appearances, and biographical information from any unauthorized commercial use. Or Congress could take a narrower approach focused only on particular identity attributes, like name and likeness. Congress would also need to determine whether a federal right-of-publicity statute preempts state right-of-publicity laws or sets a floor that would allow state protections to exceed the federal standards.

4.              Bargaining for the use of likenesses

A federal right of publicity could present an interesting way out of the current dispute between rightsholders and AI producers. Most of the foregoing comment attempts to pull apart different pieces of potential infringement actions, but such actions are only necessary, obviously, if a mutually beneficial agreement cannot be struck between creators and AI producers. The main issue at hand is that, given the vast amount of content necessary to train an AI system, it could be financially impractical for even the largest AI firms to license all the necessary content. Even if the comments above are correct, and fair use is not available, it could very well be the case that AI producers will not license very much content, possibly relying on public-domain material, and choosing to license only a very small selection.

Something like a “right of publicity,” or an equivalent agreement between creators and AI producers, could provide alternative licensing and monetization strategies that encourage cooperation between the parties. If creators had the opportunity to opt into the use of their likeness (or the relevant equivalent for the sort of AI system in question), the creators could generate revenue when the AI system actually uses the results of processing their content. Thus, the producers would not need to license content that contributes an unknown and possibly de minimis value to their systems, and would only need to pay for individual instances of use.

Indeed, in this respect, we are already beginning to see some experimentation with business models. The licensing of celebrity likenesses for Meta’s new AI chatbots highlights an emerging opportunity for creators to monetize their brand through contractual agreements that grant usage rights to tech companies that commercialize conversational AI.[109] As this technology matures, there will be more opportunities for collaborations between AI producers—who are eager to leverage reputable and recognizable personalities—and celebrities or influencers seeking new income streams.

As noted, much of the opportunity for creators and AI producers to reach these agreements will depend on how rights are assigned.[110] It may be the case that a “right of publicity” is not necessary to make this sort of bargaining happen, as creators could—at least theoretically—pursue litigation on a state-by-state basis. This disparate-litigation strategy could deter many creators, however, and it could also be the case that a single federal standard outlining a minimal property right in “publicity” could help to facilitate bargaining.

Conclusion

The advent of generative AI systems presents complex new public-policy challenges centered on the intersection of technology and copyright law. As the Copyright Office’s inquiry recognizes, there are open questions around the legal status of AI-training data, the attribution of AI outputs, and infringement liability, which all require thoughtful analysis.

Ultimately, maintaining incentives for human creativity, while also allowing AI systems to flourish, will require compromise and cooperation between stakeholders. Rather than an outright ban on the unauthorized use of copyrighted works for training data, a licensing market that enables access to a large corpora could emerge. Rightsholders may need to accept changes to how they typically license content. In exchange, AI producers will have to consider how they can share the benefit of their use of protected works with creators.

Copyright law retains flexibility to adapt to new technologies, as past reforms reacting to photography, sound recordings, software, and the internet all demonstrate. With careful balancing of interests, appropriate limitations, and respect for constitutional bounds, copyright can continue to promote the progress of science and the useful arts even in the age of artificial intelligence. This inquiry marks a constructive starting point, although ongoing reassessment will likely be needed as generative AI capabilities continue to advance rapidly.

[1] Artificial Intelligence and Copyright, Notice of Inquiry and Request for Comments, U.S. Copyright Office, Library of Congress (Aug. 30, 2023) [hereinafter “NOI”].

^[2] Tim Sweeney (@TimSweeneyEpic), Twitter (Jan. 15, 2023, 3:35 AM), https://twitter.com/timsweeneyepic/status/1614541807064608768?s=46&t=0MH_nl5w4PJJl46J2ZT0Dw.

^[3] Pulitzer Prize Winner and Other Authors Accuse OpenAI of Misusing Their Writing, Competition Policy International (Sep. 11, 2023), https://www.pymnts.com/cpi_posts/pulitzer-prize-winner-and-other-authors-accuse-openai-of-misusing-their-writing; Getty Images Statement, Getty Images (Jan. 17, 2023), https://newsroom.gettyimages.com/en/getty-images/getty-images-statement.

[4] See, e.g., Anton Oleinik, What Are Neural Networks Not Good At? On Artificial Creativity, 6 Big Data & Society (2019), available at https://journals.sagepub.com/doi/full/10.1177/2053951719839433#bibr75-2053951719839433.

[5] William M. Landes & Richard A. Posner, An Economic Analysis of Copyright Law, 18 J. Legal Stud. 325 (1989).

[6] Id. at 332.

[7] Id. at 326.

[8] Id.

[9] See infra, notes 102-103 and accompanying text.

^[10] See generally R.H. Coase, The Problem of Social Cost, 3 J. L. & Econ. 1, 2 (1960).

^[11] Richard Posner, Economic Analysis of Law (Aspen 5^th ed 1998) 65, 79.

^[12] Coase, supra note 9, at 27.

^[13] Id.

^[14] Id. at 27.

^[15] Id. at 42-43.

^[16] U.S. Copyright Office, Library of Congress, supra note 1, at 14.

^[17] For more detailed discussion of GANs and Stable Diffusion see Ian Spektor, From DALL E to Stable Diffusion: How Do Text-to-image Generation Models Work?, Tryo Labs Blog (Aug. 31, 2022), https://tryolabs.com/blog/2022/08/31/from-dalle-to-stable-diffusion.

^[18] Id.

^[19] Id.

^[20] Id.

^[21] Id.

^[22] Id.

^[23] Jay Alammar, The Illustrated Stable Diffusion, Blog (Oct. 4, 2022), https://jalammar.github.io/illustrated-stable-diffusion.

^[24] Indeed, there is evidence that some models may be trained in a way that they “memorize” their training set, to at least some extent. See, e.g., Kent K. Chang, Mackenzie Cramer, Sandeep Soni, & David Bamman, Speak, Memory: An Archaeology of Books Known to ChatGPT/GPT-4, arXiv Preprint (Oct. 20, 2023), https://arxiv.org/abs/2305.00118; OpenAI LP, Comment Regarding Request for Comments on Intellectual Property Protection for Artificial Intelligence Innovation, Before the USPTO, Dep’t of Com. (2019), available at https://www.uspto.gov/sites/default/files/documents/OpenAI_RFC-84-FR-58141.pdf.

^[25] OpenAI, LP, Comment Regarding Request for Comments on Intellectual Property Protection for Artificial Intelligence, id. (emphasis added).

^[26] 17 U.S.C. § 107.

^[27] See, e.g., Blake Brittain, Meta Tells Court AI Software Does Not Violate Author Copyrights, Reuters (Sep. 19, 2023), https://www.reuters.com/legal/litigation/meta-tells-court-ai-software-does-not-violate-author-copyrights-2023-09-19; Avram Piltch, Google Wants AI Scraping to be ‘Fair Use.’ Will That Fly in Court?, Tom’s Hardware (Aug. 11, 2023), https://www.tomshardware.com/news/google-ai-scraping-as-fair-use.

^[28] 17 U.S.C. § 106.

^[29] Register of Copyrights, DMCA Section 104 Report (U.S. Copyright Office, Aug. 2001), at 108-22, available at https://www.copyright.gov/reports/studies/dmca/sec-104-report-vol-1.pdf.

^[30] Id. at 122-23.

^[31] Id. at 112 (emphasis added).

[32] Id. at 129–30.

[33] 17 U.S.C. § 107.

[34] Id.; see also Campbell v. Acuff-Rose Music Inc., 510 U.S. 569 (1994).

[35] Critically, a fair use analysis is a multi-factor test, and even within the first factor, it’s not a mandatory requirement that a use be “transformative.” It is entirely possible that a court balancing all of the factors could indeed find that training AI systems is fair use, even if it does not hold that such uses are “transformative.”

[36] Campbell, supra note 22, at 591.

[37] Authors Guild v. Google, Inc., 804 F.3d 202, 214 (2d Cir. 2015).

[38] OpenAI submission, supra note 13, at 5.

[39] Id. at 915.

[40] Id.

[41] Id.

^[42] Id. at 933-34.

^[43] Id. at 923. (emphasis added)

^[44] Id.

[45] Id. at 924.

[46] Kelly v. Arriba Soft Corp., 336 F.3d 811 (9th Cir. 2002).

[47] Id.

[48] Id. at 818.

[49] Id.

[50] Id. at 819 (“Arriba’s use of the images serves a different function than Kelly’s use—improving access to information on the internet versus artistic expression.”).

[51] The “public benefit” aspect of copyright law is reflected in the fair-use provision, 17 U.S.C. § 107. In Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 579 (1994), the Supreme Court highlighted the “social benefit” that a use may provide, depending on the first of the statute’s four fair-use factors, the “the purpose and character of the use.”

[52] Supra note 46, at 820.

[53] Perfect 10 Inc. v. Amazon.com Inc., 487 F.3d 701 (9^th Cir., 2007)

[54] Id. at 721 (“Although an image may have been created originally to serve an entertainment, aesthetic, or informative function, a search engine transforms the image into a pointer directing a user to a source of information.”).

[55] Id. at 721.

[56] Id. at 723 (emphasis added).

[57] Id. (emphasis added).

[58] Id.

[59] Supra note 37, at 218.

[60] Id. at 215-16.

[61] Id. at 214. See also id. (“The more the appropriator is using the copied material for new, transformative purposes, the more it serves copyright’s goal of enriching public knowledge and the less likely it is that the appropriation will serve as a substitute for the original or its plausible derivatives, shrinking the protected market opportunities of the copyrighted work.”).

[62] Id. at 218.

[63] Perfect 10, 487 F.3d at 721-22 (citing Kelly, 336 F.3d at 818-19). See also Campbell, 510 U.S. at 579 (“The central purpose of this investigation is to see, in Justice Story’s words, whether the new work merely ‘supersede[s] the objects’ of the original creation, or instead adds something new, with a further purpose or different character….”) (citations omitted).

[64] See supra, notes 9-14 and accompanying text.

[65] See, e.g., the development of the compulsory “mechanical royalty,” now embodied in 17 U.S.C. § 115, that was adopted in the early 20^th century as a way to make it possible for the manufacturers of player pianos to distribute sheet music playable by their instruments.

[66] See supra notes 9-14 and accompanying text.

^[67] U.S. Copyright Office, Library of Congress, supra note 1, at 15.

[68] See infra, notes at 102-103 and accompanying text.

^[69] U.S. Copyright Office, Library of Congress, supra note 1, at 15.

^[70] See, e.g., Copyright Free Music, Premium Beat By Shutterstock, https://www.premiumbeat.com/royalty-free/licensed-music; Royalty-free stock footage at your fingertips, Adobe Stock, https://stock.adobe.com/video.

^[71] U.S. Copyright Office, Library of Congress, supra note 1, at 19.

^[72] Id.

^[73] U.S. Const. art. I, § 8, cl. 8.

[74] See Ajay Agrawal, Joshua S. Gans, & Avi Goldfarb, Exploring the Impact of Artificial Intelligence: Prediction Versus Judgment, 47 Info. Econ. & Pol’y 1, 1 (2019) (“We term this process of understanding payoffs, ‘judgment’. At the moment, it is uniquely human as no machine can form those payoffs.”).

^[75] Joshua Gans, Can AI works get copyright protection? (Redux), Joshua Gans’ Newsletter (Sept. 7, 2023), https://joshuagans.substack.com/p/can-ai-works-get-copyright-protection.

^[76] See Nicholas Carlini, et al., Extracting Training Data from Diffusion Models, Cornell Univ. (Jan. 30, 2023), available at https://arxiv.org/abs/2301.13188.

^[77] See Chang, Cramer, Soni, & Bamman, supra note 24; see also Matthew Sag, Copyright Safety for Generative AI, Working Paper (May 4, 2023), available at https://ssrn.com/abstract=4438593.; Andrés Guadamuz, A Scanner Darkly: Copyright Liability and Exceptions in Artificial Intelligence Inputs and Outputs, 25-27 (Mar. 1, 2023), available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4371204.

^[78] Laureyssens v. Idea Grp. Inc., 964 F.2d 131, 140 (2d Cir. 1992), as amended (June 24, 1992).

^[79] Id. at 139.

^[80] Harney v. Sony Pictures Television Inc., 704 F.3d 173, 178 (1st Cir. 2013). This assumes, for argument’s sake, that a given model is not “memorizing,” as noted above.

^[81] Id. at 178-79.

^[82] Id.

[83] U.S. Copyright Office, Library of Congress, supra note 1, at 25.

^[84] Notably, the state of mind of the user would be irrelevant from the point of view of whether an infringement occurs. All that is required is that a plaintiff owns a valid copyright, and that the defendant infringed it. 17 U.S.C. 106. There are cases where the state of mind of the defendant will matter, however. For one, willful or recklessly indifferent infringement by a plaintiff will open the door for higher statutory damages. See, e.g., Island Software & Computer Serv., Inc. v. Microsoft Corp., 413 F.3d 257, 263 (2d Cir. 2005). For another, a case of criminal copyright infringement will require that a defendant have acted “willfully.” 17 U.S.C. § 506(a)(1) (2023), 18 U.S.C. § 2319 (2023).

^[85] Id. at 550.

[86] Legally speaking, it would be incoherent to suggest that there can be secondary liability without primary liability. The way that AI systems work, however, could prompt Congress to modify the law in order to account for the identified situation.

^[87] See, e.g., Metro-Goldwyn-Mayer Studios Inc. v. Grokster Ltd., 380 F.3d 1154, 1160 (9th Cir. 2004), vacated and remanded, 545 U.S. 913, 125 S. Ct. 2764, 162 L. Ed. 2d 781 (2005).

^[88] See BMG Rts. Mgmt. (US) LLC v. Cox Commc’ns Inc., 881 F.3d 293, 306 (4th Cir. 2018); Sony Corp. of Am. v. Universal City Studios Inc., 464 U.S. 417, 442 (1984).

^[89] A&M Recs. Inc. v. Napster Inc., 239 F.3d 1004, 1022 (9th Cir. 2001), as amended (Apr. 3, 2001), aff’d sub nom. A&M Recs. Inc. v. Napster Inc., 284 F.3d 1091 (9th Cir. 2002), and aff’d sub nom. A&M Recs. Inc. v. Napster Inc., 284 F.3d 1091 (9th Cir. 2002).

^[90] See, e.g., Content Filtering, Microsoft Ignite, available at https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/content-filter (last visited Oct. 27, 2023).

^[91] Note that, if an AI producer can demonstrate that they used no protected works in the training phase, there may in fact be no liability for infringement at all. If a protected work is never made available to the AI system, even an output very similar to that protected work might not be “substantially similar” in a legal sense.

^[92] Copyright Review Board, Second Request for Reconsideration for Refusal to Register Théâtre D’opéra Spatial (SR # 1-11743923581; Correspondence ID: 1-5T5320R), U.S. Copyright Office (Sep. 5, 2023), available at https://fingfx.thomsonreuters.com/gfx/legaldocs/byprrqkqxpe/AI%20COPYRIGHT%20REGISTRATION%20decision.pdf.

^[93] Am. Broad. Companies Inc. v. Aereo Inc., 573 U.S. 431, 453 (2014). (Thomas J, dissenting).

^[94] Id. at 451.

[95] U.S. Copyright Office, Library of Congress, supra note 1, at 21.

^[96] See 5 U.S.C. § 1051 et seq. at § 1127.

^[97] See, e.g., ETW Corp. v. Jireh Pub. Inc., 332 F.3d 915, 923 (6th Cir. 2003).

^[98] Dastar Corp. v. Twentieth Century Fox Film Corp., 539 U.S. 23, 34 (2003).

^[99] Id. at 37.

^[100] Id. at 32.

^[101] See, e.g., Multi Time Mach. Inc. v. Amazon.com Inc., 804 F.3d 930, 938 (9^th Cir. 2015); EarthCam Inc. v. OxBlue Corp., 49 F. Supp. 3d 1210, 1241 (N.D. Ga. 2014); Coll. Network Inc. v. Moore Educ. Publishers Inc., 378 F. App’x 403, 414 (5th Cir. 2010).

^[102] Digby Adler Grp. LLC v. Image Rent a Car Inc., 79 F. Supp. 3d 1095, 1102 (N.D. Cal. 2015).

^[103] J. Thomas McCarthy, The Rights of Publicity and Privacy § 1:3. Introduction—Definition and History of the Right of Publicity—Simple Definition of the Right of Publicity, 1 Rights of Publicity and Privacy § 1:3 (2d ed).

^[104] See id. at § 6:3.

^[105] Compare Ind. Code § 32-36-1-7 (covering name, voice, signature, photograph, image, likeness, distinctive appearance, gesture, or mannerism), with Ky. Rev. Stat. Ann. § 391.170 (limited to name and likeness for “public figures”).

^[106] See Restatement (Third) of Unfair Competition § 46 (1995).

^[107] See, e.g., Jeff Harleston, Artificial Intelligence and Intellectual Property – Part II: Copyright, U.S. Senate Comm. on the Judiciary Subcomm. on Intellectual Property (Jul.12, 2023), available at https://www.judiciary.senate.gov/imo/media/doc/2023-07-12_pm_-_testimony_-_harleston1.pdf; Karla Ortiz, “AI and Copyright”, U.S. Senate Comm. on the Judiciary Subcomm. on Intellectual Property (Jul. 7, 2023), available at https://www.judiciary.senate.gov/imo/media/doc/2023-07-12_pm_-_testimony_-_ortiz.pdf; Matthew Sag, “Artificial Intelligence and Intellectual Property – Part II: Copyright and Artificial Intelligence”, U.S. Senate Comm. on the Judiciary Subcomm. on Intellectual Property (Jul. 12, 2023), available at https://www.judiciary.senate.gov/imo/media/doc/2023-07-12_pm_-_testimony_-_sag.pdf.

^[108] Authors, Attribution, and Integrity: Examining Moral Rights in the United States, U.S. Copyright Office (Apr. 2019) at 117-119, https://www.copyright.gov/policy/moralrights/full-report.pdf.

[109] Benj Edwards, Meta Launches Consumer AI Chatbots with Celebrity Avatars in its Social Apps, ArsTechnica (Sep. 28, 2023), https://arstechnica.com/information-technology/2023/09/meta-launches-consumer-ai-chatbots-with-celebrity-avatars-in-its-social-apps; Max Chafkin, Meta’s New AI Buddies Aren’t Great Conversationalists, Bloomberg (Oct. 17, 2023), https://www.bloomberg.com/news/newsletters/2023-10-17/meta-s-celebrity-ai-chatbots-on-facebook-instagram-are-surreal.

[110] See supra, notes 8-14 and accompanying text.

Abstract

The EU Commission’s recent proposal for a regulation on standard essential patents (SEPs) envisages a radical overhaul of the current framework, introducing an essentiality check system, a conciliation process for fair, reasonable and non-discriminatory (FRAND) terms, and a mechanism to determine a reasonable aggregate royalty. However, both the economic justification and the approach endorsed by the proposal are questionable. Indeed, on one hand, there is no evidence of a market failure to justify the initiative and, in addition, the provisions appear to be one-sided, apparently being aimed only at addressing a hold-up problem and pursuing a value-distribution goal from SEP owners to implementers. Accordingly, this paper views the proposal critically, arguing that it departs from the well-established meaning and rationale of FRAND commitments by disregarding hold-out problems, and it jeopardises the suitability of SEPs to serve as valuable financial collateral, thereby endangering future investments in innovation.

Dear Chairman Sanders, Ranking Member Cassidy, Chairman Smith, and Ranking Member Neal:

As former judges, former government officials, and scholars who are experts in patent law, healthcare policy, or both, we write to express our concerns about lobbying efforts for the government to impose price controls on patented drugs. Some activists and academics have written to Congress and to agency officials arguing that existing laws are “tools” for the government to impose price controls on patented drugs to lower drug prices.[1] Their arguments mischaracterize these statutes by inaccurately claiming that Congress has endorsed the imposition of price controls on patented drugs. It has not.

Drug pricing presents a multi-dimensional policy issue because the U.S. healthcare system comprises a complex, intermingled system of federal and state laws and regulations, as well as a myriad of equally complex and intermingled set of public and private institutions. Yet, activists and others inaccurately reduce the causes of drug prices to a single issue: patents. They argue that the federal government can “lower drug prices by breaking patent barriers,”[2] and they claim that two statutes can be used to achieve this policy goal: the Bayh-Dole Act and 28 U.S.C. § 1498.

Neither the Bayh-Dole Act nor § 1498 are price-control statutes, and thus they do not authorize the federal government to impose price controls on patents. This is clear by their plain legal text, as well as by their consistent interpretation by courts and agencies. The Bayh-Dole Act promotes the commercialization of patented inventions that may result from government funding of research, and § 1498 secures patent-owners in obtaining compensation for unauthorized uses of their property rights by the government. Neither law says anything about drug prices. If the government used either law to impose price controls on patented drugs, this would conflict with the clear purpose of these statutes. It would also represent an unprecedented and fundamental change in U.S. patent law. From 1790 through the twentieth century, Congress rejected bills that would impose compulsory licensing on patents.[3] The calls to use the Bayh-Dole Act or § 1498 for similar purposes fundamentally are at odds with these statutes and threaten to undermine the U.S. patent system’s historic success as a driver of U.S. global leadership in biopharmaceutical innovation.

This letter explains why neither the Bayh-Dole Act nor § 1498 can be used to break patents to impose price controls on drugs. First, it sets forth the proven success of the patent system as a driver of innovation in healthcare, which is the framework to evaluate the argument to “lower drug prices by breaking patent barriers.”[4] This argument threatens to undermine the legal system that has saved lives and improved everyone’s quality of life. It then describes the Bayh-Dole Act and § 1498, explaining how neither authorizes price controls on patented drugs. The policy argument seeking to impose price controls on drugs contradicts the clear text and purpose of these statutes.

Read the full letter here.

[1] See Letter to Senator Elizabeth Warren from Amy Kapczynski, Aaron S. Kesselheim, et al., at 1 (Apr. 20, 2022), https://tinyurl.com/yt62wt4t. Professor Kapczynski and Professor Kesselheim are the co-authors of this letter, which is based on their articles, and thus this letter is identified as the “Kapczynski-Kesselheim Letter.”

[2] Id. at 8

[3] See, e.g., Bruce W. Bugbee, Genesis of American Patent and Copyright Law 143-44 (1967) (discussing the rejection of a Senate proposal for a compulsory licensing requirement in the bill that eventually became the Patent Act of 1790); Kali Murray, Constitutional Patent Law: Principles and Institutions, 93 Nebraska Law Review 901, 935-37 (2015) (discussing 1912 bill that imposed compulsory licensing on patent owners who are not manufacturing a patented invention, which received twenty-seven days of hearings, but was not enacted into law).

[4] Kapczynski-Kesselheim Letter, supra note 1, at 8.

Abstract

This paper builds on previous work addressing the problem of “hold-out” in the licensing of standards essential patents (SEPs) in mobile cellular communications technology. Given the pervasiveness of mobile technology, and the need to maintain continued innovation in such technology, the robustness of the licensing marketplace for patents is an economically important issue. We show how the ease with which implementers of such technology such as smartphone makers can use the technology without having agreed to licenses is a major structural factor that shifts bargaining power in licence negotiations towards the implementers. Together with frictions in the enforcement process, and the increasing propensity to resist licensing by new groups of implementers (i.e., “hold out”) we explain why there is an elevated risk that the licensing marketplace may produce outcomes that are inconsistent with the “balance” that Standards Development Organizations (“SDOs”) such as ETSI have sought out. The ability of the licensing marketplace to strike this balance is critical to the continued robustness of the wireless ecosystem. We explain that there is a risk that the SEP holders’ obligation to be prepared to make licences available on Fair, Reasonable and Non-Discriminatory (FRAND) terms can be used to “bound” the worst case scenario for an implementer– i.e., that it can never do worse than receiving the “FRAND” royalty. We discuss how courts and policymakers should sensibly interpret the bounds and limits of the FRAND commitment, in order to respect the overarching goals of “balance” and robust innovation in the ecosystem.

Abstract

How should a technology firm adjust resource allocation between external and internal R&D in response to stronger patent protection? External R&D provides the firm with another channel of earnings to mitigate diminishing returns to internal R&D, but yields the firm only a fraction of the additional profit generated. Theoretically, if the marginal return to external R&D diminishes more slowly than the marginal return to internal R&D, the firm should increase external R&D more than internal R&D. Exploiting regional differences in the strengthening of patent protection due to the U.S. Court of Appeals for the Federal Circuit (CAFC), we find that the CAFC was associated with 35 percent more external R&D vis-a-vis 20 percent more internal R&D. The difference was more pronounced in industries where patents were less effective in the appropriability of product inventions and among firms more specialized in technology.

Abstract

This paper uses an unusually rich 21st century data set to compare two sets of vintage bestsellers from the early 20th century that, by a circuitous path of copyright law alterations, came to have different copyright treatments. The most striking result is that, on average, copyrighted vintage bestsellers sell almost four times as many copies as public domain vintage bestsellers and this result holds throughout the sales distribution. This result conflicts with the expectation that copyright would restrict sales by allowing the exercise of monopoly power, and instead points to factors such as post-creation investment by publishers as being much more important than previously recognized. These greater sales occur despite a price premium that we find for copyrighted works, which is on average of a size similar to royalty payments typically paid to authors, although it appears to be considerably higher for better selling editions. We also find, contrary to previous claims, that vintage copyrighted titles are slightly more likely to be sold in the market than are works in the public domain. These results imply that copyright is more likely to be socially beneficial than previously thought and substituting our superior sales data into a previously published model confirms this implication. Further, these results imply that retroactive copyright extensions are socially advantageous and that indefinitely renewable copyright is more likely to be an optimal policy.

Abstract

The EC has proposed a regulatory framework for SEPs, the heart of which is the establishment of a regulatory authority—a “competence center”—charged with maintaining a registry of SEPs with detailed information drawn from required submissions by SEP holders and “administering a system for essentiality checks and processes for aggregate royalty determination and FRAND determination.” The proposal’s stated aim is to facilitate licensing negotiations between SEP holders and implementers, applying a balanced approach towards the bargaining parties. The approach is highly unbalanced, however. It would sharpen incentives for holdout by implementers and thereby substantially weaken SEP holders’ ability to appropriate the value of their IP. In particular, implementers would be empowered to substantially delay requests by SEP holders for injunctive relief against infringement in national courts of law. It is a truism that justice delayed is justice denied. Likewise, IP rights delayed are IP rights denied. Beyond delay, the Proposal would entirely bar the recovery of some losses from infringement in certain circumstances. As a result, the practical effect of the Proposal would be to induce licensing disputes where there would otherwise have been none, supplanting private bargaining with a less well-informed and inefficient administrative process that would materially depress incentives for innovation and standardization.

Abstract

Recent controversies in the District of Delaware raise questions regarding the degree to which litigants may or should be compelled to disclose ownership, funding, and management structures. The controversies have led to an extraordinary series of hearings resulting ultimately in a challenge to the authority of a district judge to investigate compliance with disclosure rules and related litigation conduct. This article surveys six different rationales for compelling disclosure. It discusses their strengths and weaknesses and the scope of disclosure they support. It concludes that increased disclosure is desirable and that most of the disclosure ordered in the current controversies in Delaware is justified under current law. The article also proposes amendments to the FRCP, Title 28, and the Patent act to facilitate future case management.

I. Introduction

On 27 April 2023, the European Commission published its Proposal for a Regulation on Standard Essential Patents (“SEP Regulation”). The proclaimed aims of the SEP Regulation are to: 1. ensure that end users, including small businesses and EU consumers, benefit from products based on the latest standardised technologies; 2. make the EU attractive for standards innovation; and 3. encourage both SEP holders and implementers to innovate in the EU, make and sell products in the EU, and be competitive in non-EU markets.[1]

We are grateful for the opportunity to provide comments on the proposed SEP Regulation in the context of public feedback. The following is a summary of our observations:

1. The available evidence does not demonstrate the existence of a market failure in SEP-licensing markets that would justify regulatory oversight. Instead, the Commission’s own evidence points to the low incidence of SEP litigation and no systemic negative effects on SEP owners and implementers. The mobile-telecommunication market—which is claimed to have the most SEP litigation and licensing inefficiencies—has over the years seen rapid growth, expansion, declining consumer prices, and new market entry.
2. Some market imperfections are necessary-but-not-sufficient conditions for regulatory intervention. Regulation might not be necessary or proportionate if its aims could be achieved with less costly instruments.
3. The proposed SEP Regulation appears to pursue the value-redistributive function of imposing costs on only one group (SEP owners), while accruing all benefits to non-EU-based standard implementers. It is difficult to find justification for such value redistribution from the evidence presented on the functioning of SEP licensing markets.
4. The proposed SEP Regulation applies to all standards licensed on FRAND terms. It is unclear how many standards will be caught and why all standards licensed on FRAND terms are presumed to be inefficient, requiring regulatory intervention. One early study identified 148 standards licensed on FRAND terms in a 2010 laptop. No evidence was presented that licensing inefficiencies of these standards caused harms in laptop markets.
5. Evaluators and conciliators need to be qualified and experienced experts in relevant fields. There are unlikely to be enough evaluators to conduct essentiality checks reliably on such a massive scale.
6. The proposed SEP Regulation raises competition concerns, as it may allow implementers to exchange sensitive commercial information that could lead to a buyers cartel.
7. Aggregate royalty-rate notifications and nonbinding expert opinions on global aggregate royalty rates may not produce meaningful inputs and may lead to even more confusion for implementers.
8. The proposed SEP Regulation has extraterritorial effects. While the SEP Register and essentiality checks apply only for patents in force in EU Member States, a nonbinding opinion on aggregate royalties and FRAND determination will be worldwide, covering portfolios in other countries.
9. Other countries may follow and introduce their own regulations on SEPs. Such regulations may be used as a strategic and protectionist tool to devaluate the royalties of innovative European SEP owners. The proliferation of regulatory regimes would make SEP licensing even more costly, with unknown effects on the viability of the current system of collaborative and open standardisation.

Considering the above, it is our view is that the proposed SEP Regulation, in its current form, is unnecessary, disproportionate, and likely to harm both European innovators and Europe’s technology leadership on a global stage.

Nevertheless, this is not to say that the SEP licensing framework cannot be further refined and simplified. It may be possible to find solutions that might improve the existing system in a cost-effective, balanced, and efficient way. We propose some private ordering instruments as an alternative to regulation, which could be used to make licensing in the Internet of Things (“IoT”) more efficient and transparent.

II. No Evidence of a Market Failure Justifying Regulatory Intervention

The current system of SEP licensing consists of bilateral negotiations and collective licensing via patent pools. The overwhelming majority of licensing agreements are concluded amicably,[2] but in cases where parties cannot agree, litigation may become necessary.[3] This is, of course, a feature of commercial disputes of all kinds.[4] Over the years, courts have proven more than capable of resolving various contentious questions about SEPs. For instance, they gave promulgated guidance regarding if and under what conditions the SEP owner can request and obtain an injunction for infringement of an SEP;[5] what the FRAND rate between the parties ought to be;[6] the scope of a FRAND license, whether global or national;[7] the meaning of a FRAND commitment’s non-discrimination requirements;[8] whether FRAND commitments require SEP owners to offer licenses at different levels of the production chain;[9] and how to adjudicate allegations of patent holdup (supposedly opportunistic behaviour of SEP owners attempting to charge more than FRAND terms) and holdout (implementers intentionally delaying or avoiding the conclusion of a licensing agreement).[10] The Court of Justice of the European Union (CJEU) provided a framework in Huawei v ZTE for good-faith license negotiation. Courts of the EU Member States have subsequently become accustomed to evaluating the conduct of both parties and have produced substantial case law and guidance on the contents of good-faith licensing negotiations.[11]

Despite successful interventions by the courts, the Commission is concerned that the current SEP licensing and litigation system is fraught with problems and inefficiencies. Three alleged major problems have been suggested as justifying regulatory intervention.

First are high transaction costs and licensing uncertainties. According to the Commission, the average per-licence bilateral negotiation costs for the SEP holder and implementer are estimated to be between €2 million and €11 million.[12] The Commissions asserts that licensing uncertainties follow from insufficient transparency of SEP ownership and essentiality, lack of information about FRAND royalties, and a dispute system not adapted for FRAND determination.[13] That system is said to be dissatisfactory for both parties.

The Commission maintains that SEP owners face long negotiations and high costs of licensing.[14] To better assess the value that the technology brings to standard implementations, an SEP owner would have to wait several years (on average, between two and four) until the standard is implemented in the market and then approach companies with an offer to license.[15] Negotiations would then ensue, taking about three years. If no agreement is reached, litigation would add another one to two and a half years.[16] During all this time, the SEP owner would not receive any royalties for use of its technology. According to the Commission, this may explain why major SEP owners usually have licenses with only 100-200 implementers with sufficiently high volumes and/or sales value that would allow for the absorption of these costs.[17] Thus, SEP owners are unable to license the whole market. High licensing and negotiation costs may reduce their income base and incentive for participation in developing new standards.[18]

On the other hand, the Commission says that implementers face uncertainty about the costs of using standards, potentially discouraging them from implementing new technologies.[19] Implementers who take a licence are also worried about being disadvantaged against their unlicensed foreign competitors.[20] Of course, licensees are worried about competitors who do not take licences—it makes no difference whether they are foreign or home-grown. But the Commission seems to have not taken into account that this holdout is not only real, but is the most egregious example of anticompetitive behaviour.

The second supposed problem is the growing IoT market that increasingly uses technological standards from the information and communications technology (“ICT”) industry.[21] IoT markets are fragmented; volumes for certain applications may be small and profit margins tight. These industries are also not familiar with SEPs. The combination of these factors is said to make SEP licensing more difficult and expensive.

The third major concern is the protection of small and midsize enterprises (“SMEs”). According to the Commission, SMEs lack the resources to negotiate with SEP owners on an equal footing or to engage in court proceedings.[22] They also do not have sufficient licensing expertise. 84% of EU-based standard implementers are SMEs, totalling about 3,192 companies.[23]

The publicly available evidence relayed by the Commission, however, does not justify any significant concern with the current SEP-licensing system, much less a concern of such magnitude to justify extensive regulatory intervention. In fact, the Commission’s study found that high transaction costs and licensing uncertainties have not led to increased litigation or systemic negative effects.[24]

First, the Commission found that the volume of SEP-litigation cases has been relatively stable in Europe, while falling in the United States but increasing in China.[25] In recent years, the share of declared SEPs subject to litigation has declined.[26] They further showed that the prevalence of SEP litigation is low and has not increased over time. According to the study, there are fewer than 0.05 lawsuits per-license involving major SEP licensors and patent pools.[27] Regarding the effects of the current SEP-licensing system on the incentives of SEP owners and implementers, the study found no evidence that SEP owners contribute less to standards development.[28] The econometric evidence suggests that a significant share of contributions to standards development relies on patent-related incentives, indicating the importance of preserving innovation incentives for the success of the standards-development process. On the implementation side, the study found no evidence that SEP-licensing frictions lead implementers to switch to alternative (royalty-free) standards or to have systematically depressed or delayed standards implementation.[29]

The evidence from the mobile-telecommunications market, which some believe is hindered by SEP-licensing inefficiencies, demonstrates that it is functioning particularly well, with year-to-year increased output, lower prices, increased market entry, and billions of euros of investment in research and development (R&D) for connectivity standards and the rollout of new network infrastructure.^[30] For example, the latest estimate for the mobile economy in 2022 was 8.4 billion SIM connections and 4.4 billion mobile-internet subscribers, contributing $5.2 trillion or 5% of global gross domestic product, and directly and indirectly supporting 28 million jobs.^[31] In Europe, subscriber penetration was 90%, and smartphone adoption was 81%.^[32] By 2035, the impact of 5G is projected to grow to $13.2 trillion in global economic output, and the global 5G value chain will generate $3.6 trillion in economic output.^[33] Moreover, 5G is expected to add up to €1 trillion to European GDP by 2025.[34] In comparison, the total estimated revenue from cellular SEP licensing was estimated to be less than 0.5% of the mobile economy.^[35] Other studies found that the cumulative royalty yield of 2G, 3G, and 4G SEPs was only 3.4% of the smartphone’s average selling price, or just $9.60.^[36]

As to potential licensing problems in the IoT, we have yet to see the full implementation of ICT standards and corresponding SEP licensing. As such, it is too early to conclude with a sufficient degree of certainty whether there will be a systemic problem with IoT licensing. The Commission’s Impact Assessment did not provide information on the current SEP-licensing revenues obtained from various IoT sectors.[37] Thus, we do not know the current magnitude of SEP licensing in the IoT. What is certain is that IoT devices will grow in the future. According to the CRA study, cellular IoT devices represented only 20% of mobile phones in 2022, which is expected to grow to 60% by 2030.[38] As such, while licensing in the IoT may generate significant revenues, we do not at the moment have sufficient information on how many IoT devices are currently licensed.

We may observe, however, that market actors are adapting to the challenges posed by IoT. Avanci is a platform for licensing 3G, 4G, and soon 5G in the IoT.[39] It has a licensing programme for car manufacturers and has more than 120 million licensed connected vehicles.[40] Avanci includes 56 licensors and has brought together the largest SEP owners, such as Samsung, Qualcomm, Nokia, and Ericsson. It offers a one-stop solution for vehicle manufacturers with a single per unit-license of $20 per vehicle—less than a parking ticket. According to some estimates, Avanci successfully covers more than 80% of the market.[41] It may be said that SEP licensing in the automotive sector has been successfully concluded, despite the initial reluctance of car manufacturers and disputes about the appropriate level of licensing.[42]

In another example, Sisvel, a patent-pool administrator, experimented with a novel payment mechanism to address concerns that companies that take a license are disadvantaged against their unlicensed competitors. For its Wi-Fi 6 pool, it provided a licensing programme that adjusts royalty payments based on the percentage of the licensed market.[43] In other words, most royalty payments will be deferred, unless and until other competitors also pay. Such a mechanism protects licensees from patent-infringement liability, while paying only a fraction of the due royalties until the whole market takes a licence. The experience of the Avanci and Sisvel pools demonstrates that SEP owners are adjusting to the changed market realities and looking for ways to simplify licensing, with innovative structures to address the need for certainty and transparency in the IoT.

As to the supposed harmful impact of the current system of SEP licensing on SMEs, it is difficult to draw such a conclusion from the available evidence. The Commission noted that most SMEs are de facto unlicensed because licensing costs outweigh potential licensing revenues.[44] To better understand the views of SMEs, the Commission carried out two surveys—a general one in which all stakeholders could participate and a targeted one only for SMEs. The Commission received responses from nine SMEs in the general survey, while 37 SMEs participated in the targeted survey.[45] That represents a sample of only 1.15% of the 3,192 SMEs that are reported to implement standards, making it impossible to draw general conclusions from such a limited sample. The question may be asked: if SMEs face licensing problems, why have they not expressed more interest in surveys? The only answer one can reasonably draw is that there is no problem. The SME survey shows some licensing; seven out of 37 SMEs had a license.[46] It would be interesting to know, however, which SEP owners approached and licensed SMEs, as well as the licensing policies of major SEP owners toward SMEs. We do not currently possess such information.

While there is no evidence that the current SEP licensing framework has produced systematic negative effects, this is not to say that the system could not be improved. Evidence still shows that licensing costs are not insignificant and that it takes years to conclude licensing agreements. Moreover, it is unlikely that every SEP owner could reach every implementer in the IoT, thus creating an uneven playing field between licensed and unlicensed implementers.

It is likely possible to improve the existing system in a cost-effective, balanced, and efficient way, including through private and public ordering instruments.[47] If the aims could be achieved with less costly instruments, extensive regulatory intervention might be neither necessary nor proportionate.[48] In other words, the existence of market imperfections is necessary but not sufficient conditions for regulatory intervention. Regulators should also be mindful not to fall into the “nirvana fallacy”, striving for ideal but unrealistic solutions that produce more costs than other feasible alternatives that may not lead to ideal results.[49]

III. Evaluating the Effects of SEP Regulation on SEP-Licensing Markets

While the Regulation pursues the worthwhile goals of increasing transparency and certainty to parties in SEP licensing, it is improbable that the proposed solutions will achieve those aims. This section raises several issues that should be considered in future policy discussions.

A. The Regulation’s Value-Redistributive Function

The Regulation imposes unbalanced costs and benefits. According to the Impact Assessment, SEP owners will bear all the costs, while implementers will reap all the benefits.[50] The 10-year average approximate annual benefits for SEP implementers are estimated to be €24.4 million, while for SEP owners, the costs are €28.9 million. As such, the Regulation does not attempt to improve conditions for all actors (i.e., pursue Pareto efficiency) but directly seeks to redistribute value from SEP owners to implementers. The Commission notes that a large part of SEP owners’ costs would be due to an expected increase in patent fees, thanks to the anticipated rise in the number of patents.[51] It adds that patent fees would represent revenue to European and national patent offices, making the whole system socially profitable.

The Commission recognised that it is difficult to predict the impact of SEP Regulation on royalty level. The Regulation’s effects may go in two opposing directions: 1. potentially more firms taking a license (increasing implementation costs and income for SEP owners), or 2. potentially lower royalties paid (decreasing implementers’ costs and SEP owners’ income).[52] The latter scenario would place even more costs on SEP owners. If royalty revenues fall and licensing costs of increase, an unintended but obvious consequence could be that SEP owners may no longer find collaborative standardisation attractive and might instead pursue proprietary solutions unencumbered by FRAND commitments. A fragmented global system would surely impede innovation.

The EU-based implementers will not even be among the primary beneficiaries of the Regulation’s value-redistribution.  According to the Commission’s Impact Assessment, just 8% of potential manufacturing firms are in the EU. In other words, 92% of implementers are non-EU companies. The Regulation would effectively subsidise non-EU implementers while, at the same time, harming European technology developers and Europe’s technological leadership.

It is difficult to see justifications for such value redistribution from the evidence presented on the functioning of SEP licensing. In our view, any regulation should attempt to lead to better outcomes than the perceived harms it seeks to address.

B. The Regulation’s Broad Scope

The Regulation has a very broad scope and applies to an unknown number of standards. Once it enters into force, the Regulation would catch all FRAND-committed SEPs.[53] It is not unclear why such broad scope is necessary. Concerns about SEP-licensing problems have focused overwhelmingly on just a few standards, mainly in cellular communication (3G, 4G, 5G) and Wi-Fi. Other standards licensed on FRAND terms have not been mentioned as potentially problematic. Nevertheless, the Regulation will apply to all standards licensed on FRAND terms.

The Commission noted that there were about 75,000 patent families of declared SEPs worldwide in 2021.[54] But we still lack information on how many standard developing organisations (“SDOs”) were analysed, nor the number of standards expected to be caught. An earlier 2010 study identified 251 technical-interoperability standards in a modern laptop, with 148 of those licensed under FRAND terms.[55] It is unclear why these 148 standards should be regulated, nor what market failures have been associated with them. If anything, a better understanding of the SEP-licensing system in the laptop market is required before introducing regulations.

The Regulation offers some exceptions from its full application for a few standards deemed unproblematic. By a special act, the Commission will designate standards and use cases “where there is sufficient evidence that … SEP licensing negotiations on FRAND terms do not give rise to significant difficulties or inefficiencies affecting the functioning of the internal market”.[56] In other words, there is a presumption that all standards with FRAND-licensing conditions are inefficient and affect the internal market’s functioning, with the onus placed on stakeholders to rebut this presumption.

Even for such unproblematic standards, the exceptions are limited; only the provisions on conciliators facilitating the agreement on aggregate royalty rates, the nonbinding expert opinion on global aggregate royalty rates, and the mandatory FRAND determination will not apply.[57] The costliest obligations—i.e., the registration of SEP and annual essentiality checks—will continue to apply even for these standards.

C. The Need for Qualified-Expert Evaluators and Conciliators

The extent of the Regulation’s reliability will depend on having qualified experts to work as evaluators and conciliators. Evaluators will need specialised knowledge of the particular technological area in which they will conduct essentiality checks. The Commission estimates that there are about 1,500 experts (650 patent attorneys and 800 patent examiners) qualified to do essentiality checks in the EU.[58]

The sheer magnitude of the task, however, will require many more evaluators and it is very doubtful that the optimal number of potential qualified experts are even available to join this process. For certain, special arrangements would need to be made with patent offices to grant patent examiners leave to conduct essentiality checks. Each year, evaluators will need to test a random sample of up to 100 SEPs if requested by each SEP owner or an implementer per standard. Thus, the amount of work may exponentially increase depending on how many standards are caught by the Regulation.

If 148 FRAND-licensed standards per laptop are to serve as a rough proxy, then we might expect more than 100-200 standards to be checked for essentiality every year. In addition, if SEP owners and implementers regularly use the possibility of testing up to 100 SEPs per standard and per SEP owner, the sheer magnitude of work may exceed the capacity of patent attorneys. Patent attorneys may find it challenging to regularly engage in such high volumes of essentiality checks while also serving other clients. And why should they do it at all unless the rate of pay is at least what they could earn in a patent law firm? To be blunt, the work would not be as much fun as acting for real clients, so the pay would probably have to be even higher to attract applicants.

Consequently, it is very unlikely that the capability even exists to annually perform a large number of essentiality checks of registered SEPs. If the requirements to become an evaluator were relaxed to address this workload, this would cast doubt on the reliability of the whole system. There is no point in building a battleship unless you are sure you can get a competent crew.

Additionally, the patent attorneys who most apt to be familiar with these technologies may well also find themselves with conflicts of interest. They will probably have worked for some SEP owners or implementers. Elaborate rules to avoid such conflicts would need to be implemented to prevent patent attorneys who were, or still are, engaged with certain clients from becoming evaluators of those clients’ registered SEPs. The conflicts problem would, of course, apply not just to individual attorneys but to their entire firms.

Conciliators would also need to be experts in the field. They might come from the ranks of retired judges, seasoned former company officials, or experienced lawyers. Conflict-of-interest provisions should also ensure their independence and impartiality in mandatory FRAND determinations.  But the job would, again, have to be sufficiently attractive, both in remuneration and in work content and culture. The Commission has made no investigation as to whether a sufficiently large pool of credible individuals could be found to make the system work.

Of course, there are well-established voluntary systems of conciliators and mediators, some of which are used now to help resolve FRAND disputes. But the proposal adds the idea of compulsory mediation or conciliation. There is scant evidence that either system works in other commercial disputes around the world, and it is unclear why it should be assumed to work here.

D. Competitive and Practical Concerns with Aggregate Royalty Rates

The Regulation also raises potential competition concerns. The participation of implementers in the process of providing expert opinion on global aggregate rates could be used as a vehicle for a buyers cartel and could devalue FRAND royalty rates. Namely, it is unclear from the text of the Regulation if implementers will be allowed to coordinate their submissions to conciliators. If this is permitted, implementers could use the process to exchange commercially sensitive information and agree on the maximum global aggregate royalties they would pay. This would be tantamount to a buyers cartel, with price fixing of input costs. Even if such coordination is not allowed, by individually submitting their maximum royalty expectations—which are made with the goal of minimising input costs—implementers might attempt to devalue SEP royalties. Given that there are far more implementers than there are SEP owners, implementers might have an outsized influence on conciliators preparing expert opinions. The Regulation also lacks competition safeguards against the exchange of commercially sensitive information by SEP owners in the process of joint notification of aggregate royalty rates, which establish the value that devices derive from using the standardised technologies in question.

Moreover, from a practical perspective, the provisions’ usefulness is questionable. The Regulation appears to allow multiple groups of SEP owners to jointly notify their views. This may add even more confusion to standard implementers. For example, some SEP owners could announce an aggregate rate of $10 per product, another 5% of the end-product price, while a third group would prefer a lower $1 per-product rate. Moreover, it is unclear what difference the joint aggregate royalty-rate notifications would bring to the existing practice of unilateral announcement of licensing terms. Many SEP owners already publicly announce their royalty programmes in advance, which was recognised by the Commission’s studies.[59] To be on the safe side, SEP owners may simply notify their maximum preference, knowing that negotiations would lead to different prices depending on the unique details of various licensees. As a result, the aggregate royalty rates may not produce meaningful data points.

Nonbinding expert opinions on global aggregate royalty rates could also add to the confusion. Implementers would likely initiate the process, which would then exist in parallel with SEP owners’ joint notifications of aggregate rates. All these different and possibly conflicting estimates might lead to even greater uncertainty. Moreover, if those providing nonbinding opinions are not universally regarded as “experts”, the parties are unlikely to respect such opinions.

Aggregate royalty notifications and nonbinding opinions might be used in the top-down method for FRAND-royalty determinations. A top-down method provides that the SEP owner should receive a proportional share of a standard’s total aggregate royalty. It requires: 1. establishing a cumulative royalty for a standard; and then 2. calculating the share in the total royalty for an individual SEP owner. This may be the reason for having aggregate royalty-rate notifications and opinions. At the same time, essentiality checks are still needed to filter out which patents are truly essential, and to assess each individual SEP owner’s share.

We caution strongly against relying too much on the top-down approach for FRAND-royalty determinations. It is not used in commercial-licensing negotiations, and courts have frequently rejected its application. Industry practice is to use comparable licensing agreements. The top-down approach was applied in Unwired Planet v Huawei only as a cross-check for the rates derived from comparable agreements.[60] TCL v Ericsson relied on this method, but was vacated on appeal.[61] The most recent Interdigital v Lenovo judgment considered and rejected its use, finding “no value in Interdigital’s Top-Down cross-check in any of its guises”.[62] Moreover, the top-down approach, as currently applied, relies only on patent counting. It does not consider that not every patent has the same value, nor that some patents may be invalid or not infringed by a specific device. Crucially, the top-down approach and aggregate royalty notifications/opinions would be related to global FRAND royalties, while the registration of SEPs and corresponding essentiality checks are limited only to EU SEPs. In other words, the SEP Regulation has extraterritorial effects, the consequences of which are discussed below.

E. Circumventing the Regulation by Litigating Outside the EU

As a result of the high costs imposed by the Regulation and the likely delays caused by mediation/conciliation, SEP owners may realistically decide to enforce their patents outside the EU, in such countries as the United Kingdom, the United States, China, and India—all of which have had SEP litigation. This would allow firms to avoid application of the Regulation entirely.[63] Judge Klaus Grabinski, president of the Court of Appeal of the Unified Patent Court, went out of his way to note just that at the Court’s opening ceremony in Luxembourg.[64]  In truth, the Regulation constitutes a statement of lack of faith that the new Court (or, indeed, any court) can do their job.

The evidence already shows that SEP litigation in China is rising, while the United States—historically, a major venue for SEP litigation—may see a renewed increase in cases should Europe become an unattractive option.[65] The UK is also a major forum that has witnessed important cases clarifying many aspects of FRAND licensing.

For its part, Europe has built an impressive case law in implementing the Huawei v ZTE judgment and clarified the steps in good-faith licensing negotiations, but it could be left behind in shaping global SEP-licensing practices if the Regulation serves to shift litigation to other jurisdictions.

F. The Geopolitical Effects

As currently drafted, the SEP Regulation has exterritorial effects, which may lead to unintended consequences. It applies to SEPs in force in one of the EU Member States. Such SEPs should be registered with the SEP Register and will be subject to essentiality checks. This is in accordance with the principle of territoriality.

The Regulation then provides, however, for a nonbinding expert opinion that will relate to a global royalty rate, and that FRAND determination shall concern a global FRAND license (unless otherwise specified by the parties). In other words, while SEP Register and essentiality checks apply only for patents in force in EU Member States, aggregate royalties and FRAND determination will be worldwide, covering portfolios in other countries.

This exterritoriality may lead to three effects. First, if the SEP Register and the result of essentiality checks for EU SEPs are used in global aggregate royalty and FRAND determinations, they will produce inaccurate results. Some patent owners focus on the United States and U.S.-based SDOs and do not patent as much in Europe. There may also be many SEPs in China and other Asian countries that do not have European counterparts.[66] It is a euro-centric view to assume that European SEPs are a sufficient basis to determine global aggregate and FRAND rates. The Commission’s Impact Assessment notes that the EU’s share of SEPs is only 15%, compared to the United States and South Korea’s shares of 19% and China’s 30%.[67]

Second, while it is true that standards are global and commercial practice is to license globally, it is a different matter altogether when legislation requires its institutions to adopt measures with extraterritorial effects. Conciliators determining global aggregate and FRAND rates would indirectly rule on foreign portfolios held by foreign companies. Other countries will not look on this favourably.

The third and principal unintended consequence is that other countries may introduce similar regulations and could easily justify their actions as incorporating a simple “best practice” from Europe. Imagine a situation in which similar regulations are adopted by other countries: requiring notification of national SEPs, conducting local essentiality checks, determining global aggregate royalty rates for a standard, and setting global FRAND-licensing terms. It would effectively transfer SEP disputes from courts into the hands of national regulators.

Moreover, the costs to SEP owners for enforcing SEPs would be compounded, since they would need to notify and pay for essentiality checks in multiple countries. The effects of these increased costs of SEP enforcement and licensing on innovation incentives and participation in collaborative standardisation would need to be assessed. A radically changed and fragmented SEP-licensing environment would also lead to even more uncertainty for both SEP owners and implementers.

An SEP regulation implemented by other countries might easily backfire and could be used as a strategic tool to devalue the royalties of innovative European SEP owners. China might be especially receptive to the idea of regulating SEP licensing. Jonathan Barnett has provided evidence regarding how China has strategically deployed competition and patent law to reduce royalties for SEPs held by foreign companies to the benefit of domestic manufacturers.[68] The EU has also launched a complaint before the World Trade Organization (“WTO”) against China’s practice of issuing broad anti-suit injunctions to prevent the enforcement of SEPs in other jurisdictions.[69] Instead of using competition and patent law, a regulation similar to the one proposed by the European Commission could attain the same industrial policy and protectionist aims.

Taken together, the proposed SEP Regulation makes licensing SEPs more costly, provides solutions that are likely to prove unworkable in practice, and risks countervailing measures by other countries that might be detrimental to European SEP owners and innovation.

IV. Market-Based Alternatives to the Proposed Regulation

Here, we suggest some measures as alternatives to the proposed Regulation. Consistent with the principle that extensive regulatory intervention might not be necessary or proportionate if the aims could be achieved with less-costly instruments, we believe small changes in the institutes of private ordering might improve the existing system in a cost-effective and balanced way. If regulatory action is to be pursued, however, then the application of the Regulation could be limited at first to only a few selected standards and/or use cases to tests its effects.

A. Pledges from SEP Owners Not to Assert SEPs Against SMEs

According to the Commission, most standard implementers are SMEs.[70] They are currently de facto unlicensed since the transaction costs apparently outweigh the expected licensing revenues. They will remain unlicensed until they achieve sufficient market scale for the licensing to become profitable. Nevertheless, there is some evidence that a small number of SMEs have a licence, but we do not have information on how many, or which SEP owners licensed those SMEs.[71]

The situation for SMEs is thus characterised by uncertainty. While most SMEs will not be approached for a license, a small number might still be targeted by some SEP owners. Those SMEs that took a licence would be disadvantaged compared to the unlicensed majority of SMEs. Additionally, SMEs are uncertain at what point they would be considered sufficiently large to trigger the interest of SEP owners.

A private-ordering solution could be for SEP owners to give a binding pledge not to enforce SEPs against SMEs. The Commission might investigate how much support such a measure has with SEP owners. Such a pledge could be given to relevant SDOs and made public. To avoid any doubt, a definition of an SME should also be provided. For example, the Commission considers an entity an SME if it has less than 250 employees and a turnover of no more than €50 million or a balance sheet of no more than €43 million.[72] Other definitions could also be considered. For instance, there may be successful companies in the IoT that employ less staff but generate large turnover and capture a significant share of the relevant market. In any event, a clear threshold should be set so that companies may know in advance at what point they would need to take a license and might expect to be approached by SEP owners.

The downside of binding pledges not to enforce SEPs against SMEs is that SMEs represent an important part of the market. As mentioned, 84% of standard-implementers in the EU are estimated to be SMEs. While it might not be profitable to license them individually, they may generate significant collective royalties. Thus, SEP owners would be renouncing a potentially substantial royalty income. A better option might be to consider ways to simplify and reduce the costs of licensing to SMEs, as discussed in the next proposal.

B. SME License-Purchasing Groups

One way for SMEs to get licensed simply and efficiently would be to form special license-purchasing groups (“LPGs”), as proposed by Ruud Peters et al.[73] LPGs would comprise SMEs with up to 15-20% share of the relevant market, and an LPG administrator experienced in patent licensing would take care of licensing negotiations on behalf of member SMEs. This option would simplify licencing for SMEs and reduce transaction costs for both sides. SEP owners would negotiate with just one entity and, with one license, could cover hundreds or thousands of SMEs that are not profitable to license individually. The benefits to SMEs would be that they could delegate licensing negotiations to experienced professionals and be ensured that they will receive a license on the same terms as other SMEs in the LPG.

It is important to note that this proposal differs from the licensing-negotiations groups (“LNGs”) suggested by the SEP Expert Group, which raise serious competition-law risks and may be considered a façade for buyers’ cartels among implementers.[74] In an LPG, there will be no discussion of product prices, profit margins, market share, the maximum amount of royalty, or licensing level. The tasks of the LPG administrator are only to check if an SME needs a license (i.e., if it produces standard-implementing products) and to negotiate such a license with individual SEP owners and pools based on their licensing programmes. In licensing negotiations, the LPG administrator would ensure that LPG members receive an appropriate volume discount, so that SMEs would not be disadvantaged relative to larger companies with significant volumes; guarantee that members comply with reporting obligations and royalty payments to qualify for a discounted rate for compliance; and attempt to negotiate a discount on past sales. If an SME that is a member of LPG does not accept a license negotiated by the LPG administrator, it would be considered an unwilling licensee, and the SEP owner might be able to sue and obtain an injunction in accordance with Huawei v ZTE.

Therefore, with appropriate competition safeguards and mechanisms against holdout, LPGs might be a vehicle for SMEs to receive a license in an efficient, inexpensive, and secure manner, and for SEP owners to cover the whole market, which is currently untapped because of the unprofitability of bilateral licensing with SMEs.

C. Support the Formation of IoT Patent Pools

Patent pools may be an effective solution for IoT use cases characterised by many implementers and where no-cross licensing is involved. We are already witnessing Avanci and Sisvel preparing and modelling new licensing programmes for different IoT applications. Patent pools would resolve many of the Commission’s concerns about transparency: they provide certainty that truly essential patents are included in a pool, and if many SEP owners accept the pool, it serves as a de facto aggregate royalty rate for a standard.

The Commission might explore ways to assist the creation of pools. The first step may be to initiate a dialogue with patent owners and pool administrators to understand what help they may need in setting new licensing programmes. Concrete measures could then be taken to incentivise and support pool formation. For example, a pool’s implementation costs are often substantial,[75] and the Commission might consider subsidising initial essentiality checks of patents included in a pool, which would be repaid after the pool starts generating licensing revenues.

D. Limit the Scope of the Proposed Regulation

If the Regulation is to be adopted in the present shape, which we think would be a mistake, its scope of application could be limited to only a few selected standards and/or use cases for which the Commission has evidence of licensing inefficiencies, and which would serve as a real-world test of the usefulness of new regulatory measures. In this way, we may observe in real time how regulatory measures would be applied in practice and their effects on SEP-licensing markets. After evaluating their effectiveness, the Regulation might later be expanded to include other standards where licensing inefficiencies have been identified, or it may be changed or completely repealed if the solutions proposed by the Regulation prove to be ineffective, burdensome, and costly, as we and many others predict they would be.

V. Conclusion

We would like to thank the European Commission for the opportunity to comment on the proposed SEP Regulation. We believe that the available evidence used by the Commission in preparation for this Regulation does not show the existence of market failure in SEP-licensing markets that justify  regulatory oversight. Quite the opposite, the mobile-telecommunications sector, which is alleged to be the most problematic, is seeing continuous growth, innovation, and market entry. The incidence of SEP litigation is low and has been declining over the years, with no systemic negative effects on SEP owners and implementers.

In our opinion, the proposed SEP Regulation would complicate SEP licensing even further and may alter incentives to innovate in the open-standardisation environment. It unevenly distributes all the benefits to implementers and costs to SEP owners, raising the costs of licensing even more. Its broad scope will capture all standards licensed on FRAND terms, despite not establishing with a sufficient degree of certainty that all these standards are problematic. The increased costs of enforcing SEPs may shift the litigation away from Europe to other parts of the world: the United States, United Kingdom, China, and India.

European courts have over the years have built impressive case law clarifying the contents of FRAND licenses and good-faith licensing negotiations. It would be a shame to see Europe lose its place in influencing the future SEP-licensing framework. Crucially, other countries may be inspired by the Commission’s SEP Regulation and decide to adopt similar regulatory regimes. Regulations implemented by other countries might easily backfire and be used for protectionist purposes and as a strategic tool to devalue the royalties of innovative European SEP owners. The primary beneficiaries of the Regulation might be non-EU based implementers, to the detriment of European innovators and Europe’s technological leadership.

While we believe the proposed SEP Regulation is unnecessary and disproportionate, this is not to say that the SEP-licensing framework cannot be further refined and simplified. The challenge, however, is to find solutions that improve the existing system in a cost-effective, balanced, and efficient way. We believe market-based mechanisms should be supported and sought over government regulation. It must also be emphasised that there is no one size-fits-all answer. Different solutions may be applied in different markets, and appropriate competition-law safeguards must be put in place to guarantee efficient market outcomes.

[1] European Commission, Explanatory Memorandum for Proposal for a Regulation of the European Parliament and of the Council on Standard Essential Patents and Amending Regulation (EU) 2017/1001, COM (2023) 232 Final (“Explanatory Memorandum”).

[2] Justus Baron, Pere Argue-Castells, Armandine Leonard, Tim Pohlman, & Eric Sergheraert, Empirical Assessment of Potential Challenges in SEP Licensing, European Commission (2023), p. 112.

[3] See European Commission, Impact Assessment Report Accompanying the Document Proposal for a Regulation of the European Parliament and of the Council on Standard Essential Patents and Amending Regulation (EU) 2017/1001, SWD(2023) 124 final (“Impact Assessment”) p. 26 (“about 70% of the implementers take a license without litigation according to the results from the public consultation”).

[4] Adapting Carl von Clausewitz’s aphorism: “Litigation is the continuation of negotiation by other means.”

[5] C-170/13 Huawei v ZTE, ECLI:EU:C:2015:477

[6] Unwired Planet v Huawei [2017] EWHC 711 (Pat).

[7] Sisvel v Haier, KZR 36/17 Federal Court of Justice (05 May 2020)

[8] Unwired Planet v Huawei; Huawei and ZTE v Conversant [2020] UKSC 37; Philips v Wiko, 6 U 183/16 Karlsruhe Higher Regional Court (30 October 2019); HEVC (Dolby) v MAS Elektronik, 4c O 44/18 Dusseldorf Regional Court (7 May 2020).

[9] Nokia v Daimler, 2 0 34/19, Mannheim Regional Court (18 August 2020); Sharp v Daimler, 7 O 8818/19 Munich Regional Court (10 September 2020).

[10] See, Sisvel v Haier, KZR 36/17 Federal Court of Justice (05 May 2020), 61 (that implementers should not engage in patent holdout by exploiting the structural disadvantage, which SEP holders face due to the limitation of their rights to assert patents in court); Optis v Apple [2022] EWCA Civ 1411, 115 (“Apple’s behaviour …. Could well be argued to constitute a form of hold out … while Optis’ contention … would open the door to holdout”); Ericsson v D-Link, 773 F.3d 1201, 1234 (Fed Cir 2014) (“The district court need not instruct the jury on hold-up or stacking unless the accused infringer presents actual evidence of hold-up or stacking. Certainly something more than a general argument that these phenomena are possibilities is necessary.”)

[11] An electronic database of court cases implementing Huawei v ZTE is available at: https://caselaw.4ipcouncil.com/guidance-national-courts.

[12] Impact Assessment p. 13.

[13] Id. at 17.

[14] Id. at 14.

[15] Id. at12.

[16] Id. at 12.

[17] Id.

[18] Id. at 16.

[19] Id. at 14.

[20] Id. at 16.

[21] Id. at 23.

[22] Id. at 17.

[23] Id. at 11.

[24] Baron et al., supra note 2.

[25] Id. at 109-110

[26] Id. at110

[27] Id. at 108, 112.

[28] Id. at 164.

[29] Id. at 164.

[30] For some of the voluminous literature, see: Alexander Galetovic, Stephen Haber, & Ross Levine, An Empirical Examination of Patent Holdup, 11(3) Journal of Competition Law & Economics 549 (2015); Keith Mallinson, Don’t Fix What Isn’t Broken: The Extraordinary Record of Innovation and Success in the Cellular Industry Under Existing Licensing Practices, 23 George Mason Law Review 967 (2016); David Teece, The “Tragedy of the Anticommons” Fallacy: A Law and Economics Analysis of Patent Thickets and FRAND Licensing, 32 Berkeley Technology Law Journal 1490 (2017); J. Gregory Sidak, Is Patent Holdup a Hoax, 3 Criterion Journal on Innovation 401 (2018); Alexander Galetovic, Stephen Haber, & Lew Zaretzki, Is There an Anti-Commons Tragedy in the Smartphone Industry, 32 Berkeley Technology Law Journal 1527 (2018); Daniel F. Spulber, Licensing Standard Essential Patents with FRAND Commitments: Preparing for 5G Mobile Telecommunications, 18 Colorado Technology Law Journal 79 (2020); Dirk Auer & Julian Morris, Governing the Patent Commons, 38(2) Cardozo Arts & Entertainment Law Journal 291 (2020).

[31] The Mobile Economy, GSMA (2023), available at https://www.gsma.com/mobileeconomy/wp-content/uploads/2023/03/270223-The-Mobile-Economy-2023.pdf.

[32] Ibid.

[33] The 5G Economy: How 5G Will Contribute to the Global Economy?, IHS Market (2019).

[34] The Impact of 5G on the European Economy, Accenture (Feb. 2021).

[35] Bowman Heiden, Jorge Padilla, & Ruud Peters, The Value of Standard Essential Patents and the Level of Licensing, 49(1) AIPLA Quarterly Journal 1, 5-6 (2021).

[36] Alexander Galetovic, Stephen Haber, & Lew Zaretzki, An Estimate of the Average Cumulative Royalty Yield in the World Mobile Phone Industry: Theory, Measurement and Results, 42 Telecommunications Policy 263 (2018); Keith Mallinson, Cumulative Mobile SEP Royalties (19 Aug. 2015); J. Gregory Sidak, What Aggregate Royalty Do Manufacturers of Mobile Phones Pay to License Standard-Essential Patents?, 1 Criterion Journal of Innovation 701 (2016).

[37] The Commission noted that SEP royalty payments in the mobile-telecommunications industry generate between EUR 14–18 billion per year (see Impact Assessment, supra note 3, at 9).

[38] Raphaël De Coninck, Christoph von Muellern, Samuel Zimmermann, & Kilian Müller, SEP Royalties, Investment Incentives and Total Welfare, CRA Study 2022, (2022), at 18-19.

[39] https://www.avanci.com.

[40] Avanci Vehicle 4G, https://www.avanci.com/vehicle/4g.

[41] Victoria Waldersee & Supantha Mukherjee, Automakers Tackle Patent Hurdle in Quest for In-Car Tech, Reuters (21 Sep. 2021), available at: https://www.reuters.com/business/autos-transportation/automakers-tackle-patent-hurdle-quest-in-car-tech-2022-09-21.

[42] Igor Nikolic, Injunctions Facilitate Patent Licensing Deals: Evidence from the Automotive Sector, CPI Columns Intellectual Property (Jun. 2022).

[43] LIFT: Accelerating Market Penetration and Levelling the Playing Fields, Sisvel (18 Jul. 2022), available at: https://www.sisvel.com/blog/wireless-communications/lift-levelling-the-playing-field-for-early-licensees.

[44] Impact Assessment, supra note 3, at 17.

[45] Id. at 63, 68.

[46] Impact Assessment, supra note 3, at 67; Another study found that only one out of 12 surveyed SMEs had a licence, see Joachim Henkel, Licensing Standard-Essential Patents in the IoT – A Value Chain Perspective on the Markets for Technology, 51 Research Policy 104600 (2022).

[47] Bowman Heiden & Justus Baron, A Policy Governance Framework for SEP Licensing: Assessing Private Versus Public Market Interventions (2021) available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3872493.

[48] Auer & Morris, supra note 30.

[49] Harold Demsetz, Information and Efficiency: Another Viewpoint, 12(1) The Journal of Law and Economics 1 (1969).

[50] Impact Assessment, supra note 3, at 58.

[51] Id.

[52] Id. at 50.

[53] Article 1(2) of the SEP Regulation.

[54] Impact Assessment, supra note 3, at 8.

[55] Brad Biddle, Andrew White, & Sean Woods, How Many Standards in a Laptop? (And Other Empirical Questions) (2013) available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1619440.

[56] Article 1(4) of the SEP Regulation.

[57] Article 1(3) of the SEP Regulation

[58] Impact Assessment, supra note 3, at 101.

[59] Impact Assessment, supra note 3, at 84-85.

[60] Unwired Planet v Huawei [2017] EWHC 711 (Pat).

[61] TCL v Ericsson, Case No. 8:14-cv-003410JVS-DFM (C.D. Cal. 2018); TCL v Ericsson, 943 F.3d 1360 (Fed. Cir. 2019)

[62] Interdigital v Lenovo [2023] EWHC 539 (Pat) 733.

[63] The Regulation requires that patent owners register SEPs if they want to litigate them against infringers in the courts of Member States (Article 20(1)). Patent owners may simply decide to litigate outside the EU. As a result, they do not register SEPs and completely avoid conducting essentiality checks or going into mandatory FRAND determinations.

[64] Rory O’Neil, Breaking: UPC Chief Urges EU to Rethink SEP Plan, ManagingIP (30 May 2023), available at: https://www.managingip.com/article/2bqbfr0uyrki1fniy9ou8/breaking-upc-chief-urges-eu-to-rethink-sep-plan.

[65] Baron et al., supra note 2, at 110.

[66] Florian Mueller, EU-Only SEP Register Can’t Serve as a Basis for Global FRAND Determinations: Proposed EU Regulation on Standard-Essential Patents Suffers from Incongruent Provisions, FossPatents (4 Jun. 2023), available at: http://www.fosspatents.com/2023/06/eu-only-sep-register-cant-serve-as.html.

[67] Impact Assessment, supra note 3, at 8.

[68] Jonathan Barnett, Antitrust Mercantilism: The Strategic Devaluation of Intellectual Property Right in Wireless Markets, Berkeley Journal of Law & Technology (forthcoming); see also Jeanne Suchodolski, Suzanne Harrison, & Bowman Heiden, Innovation Warfare, 22 North Carolina Journal of Law & Technology 175 (2020).

[69] DS611: China-Enforcement of Intellectual Property Rights, World Trade Organization (2022), available at: https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds611_e.htm.

[70] Impact Assessment, supra note 3, at 11 (84% of EU-based standard implementers are SMEs).

[71] Impact Assessment, supra note 3, at 67.

[72] European Commission, Recommendation of 6 May 2003 Concerning the Definition of Micro, Small and Medium-Sized Enterprises (2003) C 1422.

[73] Ruud Peters, Igor Nikolic, & Bowman Heiden, Designing SEP Licensing Negotiation Groups to Reduce Patent Holdout in 5G/IoT Markets in Jonathan Barnett & Sean O’Connor (eds), 5G and Beyond: Intellectual Property and Competition Policy in the Internet of Things (Cambridge University Press 2023).

[74] Contribution to the Debate on SEPs, Group of Experts on Licensing and Valuation of Standard Essential Patents (2021), available at: https://ec.europa.eu/docsroom/documents/45217; for commentary, see Nikolic, supra note 59.

[75] Michael Mattioli & Robert P. Merges, Measuring the Costs and Benefits of Patent Pools, 78(2) Ohio State Law Journal 281 (2017).