Rules That Learn: What Taiwan’s Competition Policy Offers a Digital Economy Reshaped by AI
I. Introduction
Artificial intelligence (AI) is reshaping the debate over competition policy in digital markets. Generative AI, which produces text, images, and other content, could disrupt established business models and patterns of competition. AI assistants and agents—tools that answer requests or carry out tasks on users’ behalf—are already changing how people browse the web, search for information, and shop online.
Several competition authorities have launched inquiries into both the development and delivery of these services. Their investigations cover both the upstream activities through which firms design and train generative AI models and the downstream applications that bring those models to users.[1] Authorities are examining how AI applications may become platforms in their own right, alongside concerns that limited access to cloud computing, semiconductor chips, training data, and technical expertise could constrain new competitors. Incumbents that control several stages of this process may favor their own AI assistants and agents, protecting established businesses and limiting competition from new services.
These developments may also shift competition away from individual products and toward integrated services, with firms increasingly competing through ecosystems of connected products and services.[2] That prospect raises two emerging concerns: Established ecosystems could exclude rival AI providers, while AI assistants and agents could themselves become new gateways through which businesses reach consumers. The result could be a new form of market concentration organized around a handful of ecosystems.[3]
The rise of AI is also testing digital-market regulations that several jurisdictions adopted only a few years ago.[4] These regimes largely targeted a small group of major technology firms, without AI specifically in view. Generative AI could alter the market conditions and business models on which those rules rest, calling their scope and underlying rationale into question.[5] Jurisdictions that adopted ex ante regulation—rules imposed in advance to address anticipated competitive harms—are therefore already reviewing whether their frameworks can accommodate AI-powered services.[6]
Indeed, traditional antitrust enforcement may gain renewed importance precisely because it can adapt to these changes. Purpose-built regulation can deliver immediate, targeted responses, but its specificity can also leave it vulnerable to technological change. Competition law generally moves more slowly, yet its broader standards allow authorities to assess new business practices as markets evolve.
The European Commission’s first review of the Digital Markets Act (DMA) illustrates this tradeoff.[7] The Commission undertook a consultation on whether the DMA could adequately address AI-powered services and whether it should revise the law’s list of “core platform services”—the categories of digital services subject to its requirements—or the obligations attached to them. The Commission acknowledged that the DMA cannot address every competition issue across the AI value chain. Where new conduct falls outside its obligations, traditional antitrust enforcement must therefore play a complementary role.[8]
AI may therefore change more than the distribution of market power. It could also shift competition policy away from its recent emphasis on ex ante digital-market regulation and restore a larger role for case-by-case antitrust enforcement.
This brief examines Taiwan’s alternative approach. The Taiwan Fair Trade Commission (TFTC) set out that approach in its 2023 White Paper on Competition Policy in the Digital Economy[9] and reaffirmed it in its 2026 Public Consultation Report and Policy Statement on Generative Artificial Intelligence and Competition.[10] The TFTC addresses emerging competition concerns issue by issue under existing competition law. Its cautious stance toward digital-market regulation predates the rise of generative AI, making Taiwan a useful case study of how an established antitrust framework can respond to technological change.
The TFTC’s extension of this approach to generative AI offers an opportunity to assess whether authorities can address new competitive challenges by adapting their analytical tools, rather than creating a separate regulatory regime. Its 2026 report provides a concrete basis for examining the potential advantages of ex post antitrust enforcement—intervention in response to conduct—in rapidly changing markets.
Section II of this issue brief uses the DMA as the principal benchmark to compare models of ex ante digital-market regulation and assess their adaptability to AI. Section III examines Taiwan’s approach to ex ante regulation and the framework it has developed to weigh multiple factors as market conditions change. Section III.A surveys other Asian jurisdictions to assess whether Taiwan stands apart or leads a broader regional shift in competition policy.
Section IV examines the relationship between regulation and antitrust enforcement in the AI era, asking whether this technology could restore competition law to the central role it has lost in recent digital-market policy. Section v concludes.
II. Digital-Market Regulation Meets the AI Test
Over the past five years, policymakers have turned to ex ante regulation to address the distinctive features of digital markets and the power of large online platforms. These initiatives share a premise: traditional antitrust enforcement needs additional tools to address platform conduct before competitive harm takes hold.
The DMA stands as an early and prominent example of this shift. It also reflects the European Union’s ambition to shape regulation beyond its borders through the “Brussels effect,” whereby EU standards influence business practices and policy elsewhere.[11]
The regimes differ substantially, however, in scope and design. The DMA prescribes obligations for defined service categories. The UK tailors requirements to individual firms and activities. Japan focuses on mobile ecosystems, while Germany supplements case-by-case antitrust enforcement with special powers over designated firms.
AI tests each model’s ability to adapt as services, business models, and sources of market power change. Their differences highlight the tradeoff between speed and specificity on one side and flexibility on the other. For jurisdictions still considering new rules, these experiences also underscore the continuing role of traditional competition law in addressing conduct that regulatory categories may fail to capture.
A. The DMA’s Limits in the AI Era
The Digital Markets Act identifies both substantive and procedural limits to traditional antitrust enforcement.[12] Competition rules address particular forms of market power—most notably dominance within a defined market—and specific categories of anticompetitive conduct. Enforcement also operates mainly ex post, requiring authorities to conduct detailed assessments of complex facts and economic conditions. These rules may leave gaps when firms control important gateways between businesses and consumers but do not meet the conventional threshold for dominance.
The DMA responds through a two-step framework. First, it identifies “core platform services”: categories of digital services whose economic characteristics make them particularly susceptible to weak competition and unfair practices.[13] Those characteristics may allow providers to acquire a gatekeeping position.[14] Second, the DMA designates a provider as a gatekeeper if it meets three cumulative criteria: it significantly affects the European Union’s internal market; its core platform service provides an important gateway through which businesses reach users; and it holds, or is likely soon to hold, an entrenched and durable position.[15]
Designated gatekeepers must comply with a set of obligations that largely reflect concerns European competition authorities previously addressed through antitrust enforcement. Indeed, nearly every obligation traces back to conduct by a major digital platform that authorities had already scrutinized.[16]
This architecture creates difficulties for adapting the DMA to AI. A fixed list of service categories, coupled with detailed obligations drawn from past cases, have proven ill-suited to the complexity and pace of change in digital markets.[17] To allow for future adaptation, the DMA authorizes the European Commission to investigate markets, propose additional categories of core platform services, and update gatekeeper obligations to address new practices.[18] Substantive and procedural constraints nevertheless limit these mechanisms.[19]
The nature of AI makes those limits even more consequential. Recent proceedings show that the DMA can reach AI features embedded in designated core platform services through the obligations that already apply to those services.[20] But standalone AI applications remain outside its scope if they do not fit an existing service category, regardless of their competitive significance or their provider’s market position.
The Commission’s first review of the DMA notably brought the law’s treatment of new AI entrants to the forefront.[21] Expert observers have advised the Commission that AI could strengthen existing gatekeepers or create new ones.[22] In particular, AI agents could become gateways between businesses and consumers, challenging firms that currently control access to digital markets.
Scholars have proposed two routes for adapting the DMA. Some favor market investigations to assess whether the law should add AI-related categories of core platform services.[23] Others argue that existing categories—particularly virtual assistants—are broad enough to encompass AI agents.[24]
The Commission has, at least initially, opted for a cautious approach.[25] It recognized AI’s potential as a general-purpose technology—one with applications across many sectors—while emphasizing that the technology and its uses continue to evolve.[26] Rather than immediately expand the list of core platform services, the Commission committed to monitoring AI developments and enforcing the DMA where AI forms part of a designated service or constitutes a distinct service eligible for designation.[27]
More specifically, the Commission plans to assess whether certain AI services qualify as virtual assistants and whether it should use the DMA’s existing mechanisms to update the obligations applicable to them.[28] For now, it considers the framework adequate to address existing gatekeepers’ strategies and their integration of AI into designated services.[29]
Only a few months later, however, the Commission took a more expansive approach under the Digital Services Act (DSA),[30] designating ChatGPT as a very large online search engine.[31] It described ChatGPT as a hybrid AI service that responds to users’ prompts and queries, including by searching the web. In other words, rather than classify the service by its full range of capabilities, the Commission focused on a particular function that brought it within the DSA’s definition of an online search engine.[32]
That decision raises questions about the coherence of the Commission’s approach across the two laws. Why move promptly to bring AI services within the DSA while taking a wait-and-see approach under the DMA? And because the laws define online search engines in identical terms, does ChatGPT’s DSA designation signal a broader change in how the Commission classifies AI services? In its DMA review, the Commission had instead highlighted virtual assistants as a possible category for standalone AI services.
The Commission’s work on Google’s search-data-sharing obligations adds another dimension. In that proceeding, it concluded that AI chatbots with search functions qualify to receive data under Article 6(11) of the DMA. That position suggests that the existing DMA framework can, in some circumstances, accommodate AI services that perform functions comparable to online search engines.[33]
These developments expose a broader regulatory challenge. Generative AI applications can act as platforms, connect businesses with users, and perform search functions, often within a single service. As their capabilities evolve, they will almost certainly continue to resist easy classification within established legal categories. Frameworks built around fixed categories of digital services therefore risk becoming outdated as AI applications develop new functions and business models.
B. The UK’s More Adaptable Model
A regulatory regime’s ability to keep pace with technology depends on its design and the institutions that enforce it. The United Kingdom’s approach offers greater scope than the DMA to adapt to AI and its effects on competition.[34]
Under the Digital Markets, Competition and Consumers Act (DMCCA), the Competition and Markets Authority (CMA) may designate a firm as having “strategic market status” in a digital activity if it holds substantial and entrenched market power and a position of strategic significance. The key distinction from the DMA lies in what follows designation. The UK regime allows the CMA to tailor binding requirements to each firm’s activities and business model, rather than apply a common set of obligations across designated firms.
The CMA has two principal tools. Conduct requirements govern how a designated firm must behave in the relevant digital activity. They can address existing concerns and prevent practices that exploit the firm’s position to harm consumers or businesses, or undermine fair competition.
Pro-competition interventions give the CMA broader powers to address the sources of market power. Following a market investigation, the authority may take measures to address factors that prevent, restrict, or distort competition in the designated activity. These interventions can change the underlying competitive conditions, extending beyond rules governing particular conduct. Together, the two tools give the UK regime greater adaptability than the DMA’s more prescriptive framework.
The designation process also differs. Under the DMA, the European Commission may designate a platform that meets the quantitative thresholds within 45 working days. The DMCCA instead requires a nine-month investigation, including a forward-looking assessment covering at least five years. Anush Ganesh, Jasper van den Boom, and Kena Zheng describe this contrast as the UK’s regulatory “tortoise” and the EU’s “hare.”[35] The longer UK process allows more time to assess how emerging technologies could reshape competition, including developments that authorities cannot fully anticipate when they intervene.
Whether the CMA will extend the regime beyond established technology firms to emerging AI providers remains an open question. So far, it has focused on large digital ecosystems. Its proposed requirements for Google nevertheless illustrate how the framework can respond to AI.
The CMA is considering how AI assistants and agents should affect Google’s proposed user-choice requirements, including those involving Chrome and the Android operating system. Google has made substantial changes to its search service: integrating the latest Gemini Flash model, redesigning the search box, adding conversational follow-up through AI Mode, and announcing Search agents. In response, the CMA proposes making eligibility for Google’s search choice screen—the interface through which users select a search service—technology-neutral. This would allow competing AI assistants to appear among the available options.[36]
C. Japan’s Narrower Focus on Mobile Ecosystems
Japan’s Mobile Software Competition Act (MSCA), also known as the Smartphone Act, offers a third regulatory model. The law took effect at the end of 2025 and broadly follows the DMA’s approach, imposing comparable prohibitions and obligations within a much narrower scope.[37] It targets Apple’s and Google’s control over key parts of mobile ecosystems: operating systems, app stores, smartphone browsers, and smartphone search engines.
This design may create some of the same difficulties the DMA faces in adapting to AI. But its narrower reach also limits those difficulties. Much of the digital economy falls outside the MSCA’s ex ante rules, leaving traditional competition law as the principal means of addressing competitive concerns in those markets.
The MSCA’s limited scope may therefore reduce its exposure to regulatory obsolescence. Authorities can assess new AI services and business models outside the regulated activities through case-by-case antitrust enforcement, without first fitting them into the statute’s predefined service categories.
D. Germany’s Case-by-Case Model
Germany offers a fourth model. Its 2021 competition-law reform gave the national competition authority additional powers over firms with “paramount significance for competition across markets.”[38] Section 19a of the German Competition Act allows the authority to prohibit specified practices by designated firms, which bear the burden of demonstrating that their conduct has an objective justification. The substantial overlap in the firms and practices covered has earned the provision the label “mini-DMA.”
The German model operates within competition law and illustrates how authorities can adapt that framework to new challenges. Its overlap with the DMA nevertheless raises concerns about fragmented enforcement and regulatory coherence.
Section 19a occupies an intermediate position between traditional ex post antitrust enforcement and prescriptive ex ante regulation. Shifting the burden of proof to designated firms may offer some of the speed and effectiveness associated with advance regulation. Preserving their opportunity to justify the conduct retains the flexibility of case-by-case assessment and avoids treating the specified practices as per se violations.
These models differ in scope, design, and enforcement, but AI exposes a common limitation of ex ante intervention in rapidly changing markets. Rules that prescribe conduct in advance depend on regulators’ ability to identify the sources of market power, anticipate how competition will evolve, and determine which practices could threaten contestability or fairness. Technological change can quickly unsettle those judgments by reshaping products, business models, market boundaries, and the competitive significance of individual firms. Traditional ex post antitrust enforcement offers greater scope to assess conduct as those conditions change.
E. Lessons for Jurisdictions Considering New Rules
The experience to date of digital-market regulation offers a cautionary lesson for Australia, Brazil, India, and Turkey as they consider their own regimes. Rules designed only a few years ago already face technologies policymakers did not anticipate. AI can create new services and business models while reshaping the sources and distribution of market power. Any new regime must account for that uncertainty.
In Australia, the Australian Competition and Consumer Commission (ACCC) completed a five-year inquiry into digital platform services with its final report in 2025.[39] The ACCC has examined digital platform markets since 2017 and has recommended a new ex ante competition regime, which the government supports. Its final report proposed UK-style mandatory, service-specific codes of conduct to prevent competitive harms through flexible, targeted rules. Although the ACCC urged timely legislative action, the government has yet to introduce a bill.
The report notably took a more cautious position on AI. It raised concerns about major platforms integrating generative AI into their core products and services, but acknowledged that the technology’s competitive effects remain uncertain. The report concludes that further analysis is necessary to assess the nature and intensity of competition in the sector.[40] Given AI’s rapid development and potential use throughout the economy, the ACCC emphasized that any new regime must remain flexible enough to monitor emerging technologies and their effects on digital platform competition.[41]
Brazil has moved further toward legislation. Bill No. 4,675/2025 advanced under an urgency procedure and through a revised committee report in 2026, but subsequently stalled amid disagreements among party leaders and broader trade tensions with the United States.[42] Drawing principally on the UK model, the bill would amend Brazil’s competition law and create a Digital Markets Superintendency within the Administrative Council for Economic Defense (CADE), the national competition authority. The new body could designate firms as having “systemic relevance” in digital markets for up to 10 years and impose tailored obligations drawn from a nonexhaustive statutory list.
Turkey’s proposed competition-law amendment has remained stalled since 2022. It broadly follows the DMA, defining core platform services, providing for the designation of firms with significant market power, and prescribing an exhaustive list of obligations for those firms.
While these proposals await enactment, Brazilian and Turkish authorities continue to use existing antitrust powers to address AI-related concerns. Both have adopted interim measures to preserve rival AI assistants’ access to WhatsApp’s business interface, through which providers connect their services to the platform.[43] These actions resemble the relief granted by the European Commission and Italy’s competition authority, discussed in Section IV. The Turkish Competition Authority has also launched an inquiry into how AI is changing market structures and competition.[44]
India presents a different trajectory. Its Digital Competition Bill has effectively remained on hold within the executive branch since 2024, pending a market study of appropriate designation thresholds.[45] There has nevertheless been renewed parliamentary pressure. In August 2026, a committee called for a “shift from a reactive to a proactive, ex ante regulatory framework.”[46]
Like the DMA, the proposed Indian regime would identify core digital services and authorize the Competition Commission of India (CCI) to designate firms as “systemically significant digital enterprises” using quantitative and qualitative criteria. Designated firms would then face prescriptive obligations governing their conduct.
The CCI’s 2025 AI market study nevertheless reflects doubts about the suitability of that approach for AI-driven competition.[47] The study examines how AI is changing competition within the technology sector and across industries that use it. It identifies risks from market concentration, algorithmic collusion—coordination facilitated by algorithms—and anticompetitive conduct, alongside potential gains in innovation, efficiency, and new market opportunities.[48]
After reviewing how other jurisdictions address these concerns, the CCI favored a voluntary self-audit framework and guidance for firms deploying AI systems.[49] That approach reflects greater caution about imposing prescriptive rules while the technology and its competitive effects continue to evolve.
III. Taiwan’s Case for Adaptable Antitrust Enforcement
Taiwan offers an alternative to the regulatory models discussed above. The TFTC’s 2026 report on generative AI sets out how existing competition law can address the sector’s emerging challenges.[50] It commits to a rule-of-reason approach, weighing conduct’s procompetitive and anticompetitive effects case by case.[51]
This stance predates the emergence of generative AI. In its 2023 White Paper on Competition Policy in the Digital Economy, the TFTC distinguished ex ante regulation from competition law’s original purpose. It described regulation as addressing “the problem of market failure through replacing the ‘market’ with the ‘government’.”[52] The choice between ex ante and ex post intervention therefore turns on comparative institutional costs—“which approach incurs lower social costs”—and the government’s role in the market, expressed as the difference between “steering” and “rowing” the boat.[53]
The TFTC cautioned that any role for competition authorities in administering such regulation “should be prudently assessed.”[54] DMA-style requirements and prohibitions shift authorities from reviewing conduct after the fact to controlling it in advance. That shift raises a contested institutional question: Should competition authorities exercise these regulatory powers?[55]
The white paper also questioned approaches that infer competitive harm primarily from market structure, particularly the proposition that “bigness is badness.”[56] It notes that dominant firms have lost their positions when high prices, insufficient innovation, or other weaknesses created opportunities for rivals.[57] Those experiences support the principle that competition law should not punish firms merely for being “too big.”[58]
The TFTC instead committed to an issue-driven approach focused on “resolving problems based on each case, absorbing new knowledge under existing analysis structure, studying new analysis tools, and refining the standards of lawfulness for each kind of potential anti-competitive activity.”[59] Its legislative proposals accordingly sought targeted changes to existing competition law, preserving that framework while refining its tools.[60]
The white paper’s closing observation anticipated the challenges AI would bring:
As the digital economy features frequent innovation, short life cycles, and noticeable dynamic competition, the best enforcement principle at this stage is ‘issue-driven’.[61]
The 2026 AI report carries this approach forward. It confirms that the white paper’s principles should guide enforcement in generative AI markets.[62] It also warns against inferring unlawful conduct from market structure alone. Structural indicators should support a presumption of anticompetitive conduct only when evidence shows that market structure decisively determines the practice’s overall competitive effects.[63]
The report identifies a possible shift toward ecosystem-based competition in AI: rivalry among integrated bundles of services, alongside competition among individual products and firms.[64] It also describes markets in which firms cooperate and compete at the same time. Providers of foundation models—general-purpose models that support a range of AI applications—form alliances to secure chips, data, and computing capacity. They also partner with smaller firms and startups at different stages of the AI supply chain. Rapid technological change and uncertainty help explain these relationships.[65]
Ecosystem competition complicates several core antitrust tasks. Authorities must account for competitive benefits and harms across the different groups an ecosystem serves when defining markets, assessing market power, and evaluating justifications for conduct.[66] Strong competition among ecosystems may even call a finding of dominance into question, rather than simply reduce concern about how a dominant firm behaves.
Traditional structural indicators may miss this distinction. Market shares, concentration, and entry barriers primarily measure competition “in” a market. They may give an incomplete picture when firms compete “for” the market through innovation and a successful firm’s leading position may prove temporary.[67]
For the TFTC, these challenges strengthen the case for a rule-of-reason assessment of competition both within and among ecosystems.[68] Large firms may reinforce their positions by drawing on advantages in resources, computing capacity, and talent. At the same time, rapid advances in AI models, applications, and business models can lower barriers to creating new products, raise productivity, and support innovation and the integration of AI across different fields.[69]
Authorities must weigh these effects together. The TFTC’s framework calls for examining whether restrictions within an ecosystem entrench its position against competing ecosystems. It also asks how readily users can substitute open models for closed ones, and how that choice affects market definition, market power, and opportunities for entry. Another consideration is multi-homing—the use of multiple services—which may limit the effects of restrictions within any single ecosystem.[70]
The assessment must also test claimed efficiencies, including improvements in quality and stronger competition among ecosystems, and account for innovation when weighing competitive effects. These considerations require evidence about how particular services, restrictions, and business relationships work.
Taiwan’s approach illustrates the flexibility of ex post enforcement. Its multifactor framework can evolve as authorities gain evidence and markets change. Questions such as how readily open-source models substitute for proprietary models, or how commonly users rely on multiple services, remain variables to investigate in each case.
That flexibility matters when technology changes the competitive significance of the conduct under review. If AI markets require this degree of detailed, case-specific analysis, any ex ante regime faces a difficult task in prescribing appropriate conduct in advance, whatever its institutional design.
A. A Spectrum of Approaches Across Asia
Taiwan’s caution has parallels elsewhere in Asia. Several competition authorities within the Association of Southeast Asian Nations (ASEAN) have favored guidance and existing competition law over a catalogue of ex ante digital-market obligations. Singapore, Malaysia, and the Philippines have each examined digital platform markets and, for now, favored this approach. South Korea has stepped back from a comprehensive regime while continuing to pursue narrower, sector-specific legislation.
The region nevertheless presents a spectrum of approaches. Thailand, Vietnam, and Indonesia offer important counterexamples to this cautious trend. Across these jurisdictions, authorities confront the same tension the TFTC identifies between the predictability of advance rules and the adaptability of case-by-case enforcement.
Singapore illustrates the cautious approach. Its National AI Strategy 2.0 identifies agility as a regulatory principle, which its preference for voluntary tools reflects.[71] The Infocomm Media Development Authority (IMDA) launched the AI Verify toolkit in 2022 to help firms assess their systems against internationally recognized governance principles, without requiring them to disclose the results to regulators.[72] In late 2025, the Competition and Consumer Commission of Singapore (CCS) and IMDA launched the AI Markets Toolkit, extending voluntary self-assessment to compliance with existing competition law.[73]
CCS Chief Executive Alvin Koh has expressed similar caution about adopting a DMA-style regime. He has emphasized that the DMA’s overall effectiveness remains uncertain and that digital-market regulation must reflect local conditions. Singapore leaves open the possibility of future intervention, subject to careful calibration.[74] CCS also leads an ASEAN-wide study of AI markets, extending this emphasis on gathering evidence to the regional level.[75]
The Malaysia Competition Commission (MyCC) reached a comparable position after an 18-month review of four digital subsectors.[76] Although the review identified substantial concentration, its public summary largely recommended guidelines, coordination among agencies, transparency measures, and stronger institutional capacity. Its sole recommendation contemplating enforcement concerned parity clauses—terms that restrict businesses’ ability to offer different prices through competing channels—and would rely on existing competition law.[77]
MyCC has also proposed a central digital-economy task force to develop a regional framework aligned with ASEAN standards, rather than a Malaysian gatekeeper regime.[78] At the American Bar Association’s Spring Meeting, MyCC representatives cited limited enforcement capacity, fragmented regulation, and uncertainty about technologies such as AI as reasons for caution about ex ante rules.[79]
Senior Philippine competition officials expressed similar reservations at the same meeting.[80] A 2024 study published by the Philippine Competition Commission had left several options open, including dedicated digital-market legislation and stronger enforcement under existing competition law.[81]
South Korea presents a more mixed picture. The Korea Fair Trade Commission (KFTC) abandoned its proposed Platform Competition Promotion Act, conceived as a counterpart to the DMA, following industry opposition and diplomatic pressure. But narrower ex ante intervention remains on the agenda. Proposed amendments to the Telecommunications Business Act would require designated mobile operating-system providers to give third parties access to specified hardware and software functions so their services can work with those platforms.[82] These requirements would operate within existing sector-specific legislation and cover a much narrower range of activities than the abandoned comprehensive regime.
The KFTC’s December 2024 generative AI report leaves its broader direction open.[83] It recognizes both intense competition across the AI ecosystem and concerns about potentially anticompetitive practices. Given the sector’s dynamism, the KFTC calls for further consideration of whether authorities should address such conduct preemptively or retrospectively through changes to the existing framework.
Thailand has moved toward closer scrutiny of digital platforms. The Trade Competition Commission of Thailand (TCCT) has proposed guidelines for assessing conduct by multisided e-commerce platforms—services that connect different groups, such as sellers and buyers.[84] It has also consulted on revisions to guidelines covering digital platforms, wholesale and retail trade, and on-demand delivery services.[85]
The draft e-commerce guidelines identify a broad range of potentially unfair practices, without making them per se violations. They would apply even where a platform lacks a dominant market position. The TCCT has also established two subcommittees to strengthen oversight of digital platforms and modern wholesale and retail businesses.[86]
Thailand is separately considering a Digital Platform Economy Bill modeled closely on the DMA, which officials had expected to submit to Parliament in 2026.[87] In August, the minister of digital economy and society slowed the process and ordered revisions. The minister considered the EU-inspired approach too stringent and called for rules better suited to Thai conditions.[88]
Vietnam’s proposals combine elements of digital-market regulation with conventional competition-law enforcement. The Ministry of Industry and Trade has proposed platform-specific methods for calculating market shares and additional factors for assessing market power, including data collection and user scale. The proposals also target self-preferencing, restrictions on switching, denial of access to users’ transaction data, and preferential access to commercially sensitive data.[89]
These restrictions would approach per se prohibitions in one important respect. Authorities would still need to establish foreclosure and competitive harm, but dominant firms could not invoke an efficiency defense. The rules would nevertheless operate ex post within an abuse-of-dominance framework and apply across the economy to dominant firms, rather than only to designated digital platforms.
Vietnam thus offers a hybrid model, importing substantive features of digital-market regulation into traditional competition law. It also differs from Germany’s approach, which uses a special designation process and shifts the burden to designated firms to justify specified conduct.
Indonesia’s plans remain less developed, but its competition authority has called for a dedicated Digital Markets Law. The chair of the Indonesian Competition Commission, known by its Indonesian abbreviation KPPU, cited the complexity of digital platforms, algorithmic self-preferencing, and AI as reasons to adopt a DMA-style framework.[90]
These approaches do not reveal a settled regional consensus, but they do show that authorities can scrutinize AI and digital platforms through several institutional routes. The central question remains how much conduct policymakers should prescribe in advance and how much they should assess through enforcement as evidence and markets develop.
IV. Antitrust’s Adaptive Advantage
AI is testing the assumptions behind the recent wave of digital-market regulation. Regimes adopted over the past five years sought to address market structures that policymakers viewed as entrenched and resistant to competition. A technology capable of creating new services, changing how users reach businesses, and challenging established firms could unsettle the assumptions about durable market power that was seen to justify those rules.
This uncertainty strengthens the comparative case for traditional antitrust enforcement. Ex ante regimes translate predictions about competitive problems into predefined obligations. Their effectiveness depends on regulators identifying the sources of market power, anticipating how competition will develop, and determining which practices could threaten contestability or fairness. Rapid technological change makes those judgments especially difficult.
AI illustrates the problem. It is changing competitive relationships, creating new gateways between businesses and consumers, and opening opportunities to challenge firms that appeared entrenched when policymakers designed the rules. It may also produce conduct that regulators could not reasonably have anticipated. Existing obligations could consequently miss emerging harms or restrict practices that changing market conditions have made benign or even beneficial.
Detailed rules also create opportunities for circumvention. Firms may satisfy formal requirements while pursuing alternative strategies with similar competitive effects. A catalogue of obligations cannot anticipate every response, particularly when firms have strong incentives and the technical capacity to redesign products and commercial practices around regulatory constraints.
Keeping such rules effective requires continuous monitoring, interpretation, and revision. Those efforts impose costs and can erode a principal claimed advantage of ex ante regulation—greater predictability and legal certainty. Because revisions generally follow market developments, authorities face a persistent tension between stable rules and adaptable enforcement.
Traditional antitrust has its own costs, including lengthy proceedings, complex analysis, and uncertainty. Its case-specific approach nevertheless offers a significant advantage in changing markets.[91] Authorities can assess conduct against the competitive conditions in which it occurs, adapting their analysis as products, business models, and market boundaries evolve.
Taiwan illustrates how this flexibility can work. The TFTC treats technological and market uncertainty as factors to investigate in each case. Its rule-of-reason framework can weigh competition within and among AI ecosystems while accounting for the pressure new entrants exert on established firms.
The reason that scope matters is that a framework organized around designated technology incumbents may struggle both to classify AI applications that become new gateways and to address new providers that challenge existing firms or acquire gatekeeping power themselves. Case-by-case antitrust analysis can examine these developments from the outset, without first fitting them into predetermined service categories.
The European Commission acknowledged this continuing role for antitrust in its first DMA review. It recognized that the DMA cannot address every competition concern across the AI value chain and that competition law must complement the regulation when novel conduct falls outside its obligations.[92] The Commission cited two antitrust investigations: one into Google’s use of online content for AI services and another into Meta’s restrictions on competing AI assistants’ access to WhatsApp.[93]
The Meta AI proceedings also demonstrate that antitrust need not always await the conclusion of a lengthy investigation. Interim measures allow authorities to act quickly to preserve competition while they assess the merits.[94] To prevent serious and potentially irreparable harm in the emerging market for AI assistants, the Commission and the Italian Competition Authority ordered Meta to restore rival assistants’ access to WhatsApp’s business interface on the terms that applied before its policy change.[95]
These proceedings suggest a role for competition law that extends beyond filling gaps in the DMA. Antitrust can address conduct that policymakers could not anticipate and generate evidence about how new markets work. The Commission has indicated that it will draw on its investigations when assessing the need for additional regulation.
Ex post enforcement can therefore serve two connected purposes. It can respond to emerging competitive concerns and build the knowledge needed to determine whether subsequent ex ante intervention is justified—and what form it should take.
V. Conclusion
AI is testing the assumptions behind the recent turn toward digital-market regulation. Rules designed around established gatekeepers and familiar business practices must now accommodate new services, emerging intermediaries, and competition among ecosystems. Fixed service categories and prescriptive obligations risk missing new harms while constraining conduct that changing market conditions have made beneficial.
Taiwan demonstrates how competition law can adapt to this uncertainty. The TFTC’s rule-of-reason framework examines competition within and among ecosystems, weighs efficiencies and innovation, and accounts for challenges from new entrants. It treats market structure as evidence to assess, rather than a sufficient basis for inferring unlawful conduct. This approach builds on Taiwan’s longstanding preference for refining existing antitrust tools as new problems arise.
Other Asian jurisdictions show varying degrees of caution, alongside proposals for broader intervention. The differences among the DMA, the UK’s tailored requirements, Japan’s narrower rules, and Germany’s special enforcement powers are also relevant to future comparative inquiries. Policymakers must assess each model’s institutional costs and capacity to adapt to local conditions. Frequent revisions may improve responsiveness, but they also weaken the predictability that advance regulation promises.
Antitrust enforcement carries costs of delay, complexity, and uncertainty. Yet interim measures can preserve competition while investigations proceed, and case-specific analysis can generate the evidence needed to assess future regulation. The European Commission’s AI investigations illustrate both functions.
AI therefore warrants a reconsideration of the presumption favoring ex ante digital-market rules. Taiwan offers a practical alternative: investigate concrete concerns, assess competitive effects, and adapt enforcement as evidence develops. That approach preserves room for innovation while building a firmer basis for deciding whether additional regulation is necessary.
[1] See, e.g., Autorité de la Concurrence, Conversational Agents: The Autorité Starts Inquiries Ex Officio with a View to Issuing an Opinion (Jan. 9, 2026), https://www.autoritedelaconcurrence.fr/en/press-release/conversational-agents-autorite-starts-inquiries-ex-officio-view-issuing-opinion; Japan Fair Trade Comm’n, Report Regarding Generative AI Version. 2.0 (Apr. 16, 2026), https://www.jftc.go.jp/en/pressreleases/yearly-2026/April/260416.html; Taiwan Fair Trade Comm’n, Public Consultation Report and Policy Statement on Generative Artificial Intelligence and Competition (2026), https://www.ftc.gov.tw/internet/english/doc/docDetail.aspx?uid=179&docid=18369 [hereinafter Taiwan AI Report]; Autorité de la Concurrence, Opinion on the Competitive Functioning of the Generative Artificial Intelligence Sector, Opinion No. 24-A-05 (June 28, 2024), https://www.autoritedelaconcurrence.fr/sites/default/files/commitments/2024-09/24a05_eng.pdf; Competition & Mkts. Auth., AI Foundation Models: Update Paper (Apr. 11, 2024), https://www.gov.uk/government/publications/ai-foundation-models-update-paper; Autoridade da Concorrência, Competition and Generative Artificial Intelligence (Nov. 2023), https://www.concorrencia.pt/sites/default/files/documentos/Issues%20Paper%20-%20Competition%20and%20Generative%20Artificial%20Intelligence.pdf. See also Org. for Econ. Coop. & Dev., Artificial Intelligence and Competitive Dynamics in Downstream Markets (2025), https://doi.org/10.1787/ccf0624a-en; Org. for Econ. Coop. & Dev., Competition in Artificial Intelligence Infrastructure (2025), https://doi.org/10.1787/623d1874-en.
[2] Giuseppe Colangelo & Ariel Ezrachi, AI Ecosystems, Power Shifts, and EU Competition Law Enforcement (2026), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6841798.
[3] See, e.g., Similarweb, 2026 Generative AI Landscape: The Evolution of AI Search (2026), https://www.similarweb.com/corp/reports/2026-generative-ai-landscape (showing that AI chatbots are becoming major internet destinations as platforms integrate AI into search, social media, commerce, and productivity tools that already reach billions of users).
[4] See, e.g., Philipp Hacker, Johann Cordes & Janina Rochon, Regulating Gatekeeper Artificial Intelligence and Data: Transparency, Access and Fairness Under the Digital Markets Act, the General Data Protection Regulation and Beyond, 15 Eur. J. Risk Regul. 49 (2024).
[5] See, e.g., Giuseppe Colangelo, Is AI the End of the DMA as We Know It?, GRUR Int’l (forthcoming).
[6] See, e.g., European Comm’n, Consultation on the First Review of the Digital Markets Act (July 3, 2025), https://digital-markets-act.ec.europa.eu/consultation-first-review-digital-markets-act_en [hereinafter DMA Review Consultation]; European Comm’n, Review of the Digital Markets Act: Call for Evidence for an Initiative, Ares(2025)6881572 (2025), https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=intcom:Ares(2025)6881572 [hereinafter DMA Review Call for Evidence].
[7] Regulation (EU) 2022/1925 of the European Parliament and of the Council of 14 September 2022 on Contestable and Fair Markets in the Digital Sector and Amending Directives (EU) 2019/1937 and (EU) 2020/1828 (Digital Markets Act), 2022 O.J. (L 265) 1 [hereinafter Digital Markets Act].
[8] European Comm’n, Commission Staff Working Document Accompanying the Report on the Review of Regulation (EU) 2022/1925 (Digital Markets Act), in Accordance with Article 53 Thereof, at 48, SWD(2026) 123 final (Apr. 28, 2026) [hereinafter DMA Review Staff Working Document].
[9] Taiwan Fair Trade Comm’n, White Paper on Competition Policy in the Digital Economy (Summary) (Dec. 2023), https://www.ftc.gov.tw/upload/c1697761-9974-412f-b6cb-28befbc9e8d3.pdf.
[10] Taiwan AI Report, supra note 1.
[11] Anu Bradford, The Brussels Effect: How the European Union Rules the World (2020).
[12] Digital Markets Act, recital 5.
[13] Id. recital 13.
[14] Id. recital 2.
[15] Id. art. 3(1).
[16] See, e.g., Fiona Scott Morton & Cristina Caffarra, The European Commission Digital Markets Act: A Translation, VoxEU (Jan. 5, 2021), https://cepr.org/voxeu/columns/european-commission-digital-markets-act-translation.
[17] See, e.g., Marco Cappai & Giuseppe Colangelo, Taming Digital Gatekeepers: The “More Regulatory Approach” to Antitrust Law, 41 Comput. L. & Sec. Rev. 105559 (2021).
[18] Digital Markets Act, art. 19.
[19] Id. art. 12(2).
[20] See European Comm’n, Commission Seeks Feedback on Measures to Ensure Interoperability with Google’s Android Under the Digital Markets Act (Apr. 27, 2026), https://ec.europa.eu/commission/presscorner/detail/en/ip_26_887; European Comm’n, Commission Sends Preliminary Findings to Alphabet Under the Digital Markets Act (Mar. 19, 2025), https://ec.europa.eu/commission/presscorner/detail/en/ip_25_811.
[21] DMA Review Consultation, supra note 6; DMA Review Call for Evidence, supra note 6.
[22] High-Level Grp. for the Digit. Mkts. Act, Public Statement on Artificial Intelligence (May 22, 2024), https://digital-markets-act.ec.europa.eu/high-level-group-digital-markets-act-public-statement-artificial-intelligence-2024-05-22_en.
[23] See, e.g., Thibault Schrepel, The Taxonomy Trap: AI Paradigm Shifts and the DMA (June 19, 2026), https://ssrn.com/abstract=6961084; Jan-Frederick Göhsl, Future Proofing the DMA for Agentic AI: Lessons from the AI Act, 48 World Competition 315 (2025); Ayse Gizem Yasar et al., Integration of Generative AI in the Digital Markets Act: Contestability and Fairness from a Cross-Disciplinary Perspective (LSE Law, Soc’y & Econ. Working Papers, No. 4/2024, 2024), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4769439.
[24] See, e.g., Friso Bostoen & Jan Krämer, How Future-Proof Is the DMA? A Case Study of AI Agents, J. Competition L. & Econ. (Mar. 24, 2026), https://doi.org/10.1093/joclec/nhag005; Peter Georg Picht, DMA—Innovation (2026), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6409478.
[25] European Comm’n, Report on the Review of Regulation (EU) 2022/1925 (Digital Markets Act), in Accordance with Article 53 Thereof, at 9–11, COM(2026) 178 final (Apr. 28, 2026) [hereinafter DMA Review Report].
[26] DMA Review Staff Working Document, supra note 8, § 4.2.3.
[27] DMA Review Report, supra note 25, at 11.
[28] Id.; DMA Review Staff Working Document, supra note 8, § 4.2.3.
[29] DMA Review Report, supra note 25, at 9–11.
[30] Regulation (EU) 2022/2065 of the European Parliament and of the Council of 19 October 2022 on a Single Market for Digital Services and Amending Directive 2000/31/EC (Digital Services Act), 2022 O.J. (L 277) 1 [hereinafter Digital Services Act].
[31] European Comm’n, Commission Designates ChatGPT, Reddit, Roblox Under Digital Services Act (Aug. 31, 2026), https://ec.europa.eu/commission/presscorner/detail/en/ip_26_1772.
[32] Digital Services Act, art. 3(j). Italy’s competition and communications authorities have disagreed over how to classify AI chatbots under the Act. The communications authority argues that they qualify as AI-powered search engines and thus as intermediary services within its supervisory remit as Italy’s Digital Services Coordinator. The competition authority has rejected that classification in three cases, finding that an optional, user-activated web-search function does not, by itself, make an AI chatbot an online search engine providing an intermediary service. See Autorità Garante della Concorrenza e del Mercato, DeepSeek/Informativa su “allucinazioni”, Case PS12942, Decision No. 31784 (Dec. 16, 2025); Autorità Garante della Concorrenza e del Mercato, Mistral AI/Informativa su “allucinazioni”, Case PS12968, Decision No. 31864 (Feb. 17, 2026); Autorità Garante della Concorrenza e del Mercato, Nova AI Chatbox/Informativa su “allucinazioni”, Case PS12973, Decision No. 31935 (Apr. 21, 2026).
[33] European Comm’n, Commission Provides Guidance to Google for AI Interoperability on Android and Sharing of Google Search Data Under the Digital Markets Act (July 16, 2026), https://ec.europa.eu/commission/presscorner/api/files/document/print/en/ip_26_1634/IP_26_1634_EN.pdf.
[34] Digital Markets, Competition and Consumers Act 2024, c. 13 (UK).
[35] Anush Ganesh, Jasper van den Boom & Kena Zheng, The Hare and the Tortoise: Appraising the Different Designation Procedures in EU and UK Digital Regulation, Maastricht J. Eur. & Compar. L. (Sept. 28, 2026), https://doi.org/10.1177/1023263X261487289.
[36] Competition & Mkts. Auth., Google’s General Search Services: Proposed User Choice Conduct Requirement (Sept. 23, 2026), https://www.gov.uk/government/consultations/googles-general-search-services-proposed-user-choice-conduct-requirement.
[37] Act on Promotion of Competition for Specified Smartphone Software, Act No. 58 of 2024 (Japan).
[38] Gesetz gegen Wettbewerbsbeschränkungen [GWB] [Competition Act], § 19a (Ger.) (as amended by the GWB Digitalisation Act of 2021).
[39] Austl. Competition & Consumer Comm’n, Digital Platform Services Inquiry: Final Report (Mar. 2025), https://www.accc.gov.au/about-us/publications/serial-publications/digital-platform-services-inquiry-2020-25-reports/digital-platform-services-inquiry-final-report-march-2025.
[40] Id. at 302–03.
[41] Id.
[42] Projeto de Lei No. 4.675, de 2025 [Bill No. 4,675 of 2025], Câmara dos Deputados (Braz.), https://www.camara.leg.br/proposicoesWeb/fichadetramitacao?idProposicao=2562481. See also Dario Oliveira Neto, Brazil’s Gatekeeper Bill Gets a Tuneup, Not a Rethink, Truth on the Mkt. (July 17, 2026), https://truthonthemarket.com/2026/07/17/brazils-gatekeeper-bill-gets-a-tuneup-not-a-rethink (discussing the bill’s expedited procedure, revised text, stalled floor vote, and potential trade friction with the United States).
[43] Conselho Administrativo de Defesa Econômica, Cade abre inquérito contra Meta e aplica medida preventiva suspendendo Novos Termos do WhatsApp sobre IA (Jan. 12, 2026), https://www.gov.br/cade/pt-br/assuntos/noticias/cade-abre-inquerito-contra-meta-e-aplica-medida-preventiva-suspendendo-novos-termos-do-whatsapp-sobre-ia; Rekabet Kurumu, It Was Decided to Initiate an Investigation and Impose an Interim Measure on Meta (June 5, 2026), https://www.rekabet.gov.tr/en/Guncel/it-was-decided-to-initiate-an-investigat-55485d2fcc60f11193eb0050568549fa.
[44] Rekabet Kurumu, Competition Authority Launched Sector Inquiry on AI (Apr. 7, 2026), https://www.rekabet.gov.tr/en/Guncel/competition-authority-launched-sector-in-00e7fb1b8332f11193f70050568585c9.
[45] See Comm. on Digit. Competition L., Ministry of Corp. Affs., Report of the Committee on Digital Competition Law (2024), https://www.mca.gov.in/bin/dms/getdocument?mds=gzGtvSkE3zIVhAuBe2pbow%253D%253D&type=open.
[46] Standing Comm. on Fin., Lok Sabha, Thirty-Seventh Report: Action Taken by the Government on the Observations/Recommendations Contained in the Twenty-Fifth Report on “Evolving Role of Competition Commission of India in the Economy, Particularly the Digital Landscape” 2 (Aug. 10, 2026), https://www.medianama.com/wp-content/uploads/2026/08/10Aug2026_ParliamentaryCommitteeFinanceReport.pdf.
[47] Competition Comm’n of India, Market Study on Artificial Intelligence and Competition (2025), https://www.cci.gov.in/images/marketstudie/en/market-study-on-artificial-intelligence-and-competition1759752172.pdf.
[48] Id. at 71.
[49] Id. at 99–101.
[50] Taiwan AI Report, supra note 1.
[51] Id. at 58, 61.
[52] Taiwan Fair Trade Comm’n, White Paper on Competition Policy in the Digital Economy, supra note 9, at 92.
[53] Id.
[54] Id.
[55] Id. at 22.
[56] Id. at 21.
[57] Id. at 21–23.
[58] Id.
[59] Id. at 92.
[60] Id. at 90, 92.
[61] Id. at 92.
[62] Taiwan AI Report, supra note 1, at 48.
[63] Id. at 59.
[64] Id. at 48, 56. See also Jan-Frederick Göhsl, What if Disruption Really Happens—Are Competition Law and Digital Regulation Fit for a New Era of AI-Driven Competition?, 22 J. Competition L. & Econ. 311 (2026) (arguing that incumbents protecting legacy services and successful, well-funded entrants can both engage in practices that tip AI markets and create new AI ecosystems).
[65] Taiwan AI Report, supra note 1, at 47.
[66] Id. at 48.
[67] Id. at 60.
[68] Id. at 59. See also Colangelo & Ezrachi, supra note 2.
[69] Taiwan AI Report, supra note 1, at 59.
[70] Id. at 60.
[71] Smart Nation & Digit. Gov’t Off. & Ministry of Commc’ns & Info., National Artificial Intelligence Strategy 2.0: AI for the Public Good, for Singapore and the World 54 (2023), https://file.go.gov.sg/nais2023.pdf.
[72] Ministry of Commc’ns & Info., Infocomm Media Dev. Auth. & Pers. Data Prot. Comm’n, Singapore Launches World’s First AI Testing Framework and Toolkit to Promote Transparency; Invites Companies to Pilot and Contribute to International Standards Development (2022), https://www.imda.gov.sg/resources/press-releases-factsheets-and-speeches/press-releases/2022/sg-launches-worlds-first-ai-testing-framework-and-toolkit-to-promote-transparency.
[73] Competition & Consumer Comm’n of Sing., AI Markets Toolkit (Sept. 26, 2025), https://www.ccs.gov.sg/resources/ai-markets-toolkit.
[74] Alvin Koh, The View from Singapore: A TOTM Q&A with Alvin Koh, Truth on the Mkt. (Apr. 22, 2026), https://truthonthemarket.com/2026/04/22/the-view-from-singapore-a-totm-qa-with-alvin-koh.
[75] Competition & Consumer Comm’n of Sing., International Fora (July 7, 2026), https://www.ccs.gov.sg/about-ccs/international-relations/international-fora.
[76] Malay. Competition Comm’n, Market Review on the Digital Economy Ecosystem Under the Competition Act 2010 (Feb. 10, 2026), https://www.mycc.gov.my/market-review.
[77] See also Lilien Wong & Ooi Ji Shan, Malaysia Competition Law: MyCC 2026 Digital Economy Market Review, Lexology (Mar. 12, 2026), https://www.lexology.com/library/detail.aspx?g=1cc9a743-3673-4070-812e-56591a2ab251.
[78] See Esther Lee & Jose Barrock, Special Report: MyCC Opening Digital Economy Pandora’s Box, The Edge Malay. Wkly. (Feb. 9, 2026), https://theedgemalaysia.com/node/792407.
[79] James Konstantin Galvez, Malaysia, Philippines Antitrust Regulators Hold Back on Ex-Ante Digital Rules, MLex (Mar. 26, 2026), https://www.mlex.com/mlex/articles/2457785/malaysia-philippines-antitrust-regulators-hold-back-on-ex-ante-digital-rules.
[80] Id.
[81] Jestoni A. Olivo, Digital Platforms and Online Advertising: A Guide for Competition Policy (Phil. Competition Comm’n, Market Study No. 01, 2024), https://phcc.gov.ph/file-manager/file-manager/POSTS/PCC-Market-Study-2024-01-Digital-Platforms-and-Online-Advertising_compressed.pdf.
[82] South Korea’s Telecom Business Act Amendments Target Closed Mobile OS Ecosystems, MLex (May 28, 2026), https://www.mlex.com/mlex/antitrust/articles/2482676.
[83] Korea Fair Trade Comm’n, Generative AI and Competition (Dec. 2024), https://www.ftc.go.kr/eng/selectBbsNttView.do?key=564&bordCd=822&nttSn=13613.
[84] Trade Competition Comm’n of Thai., Draft Notification on Guidelines for Considering Unfair Trade Practices and Monopolistic or Anti-Competitive Conduct in Multi-Sided Platform Businesses (E-Commerce Digital Platform Services), https://tcct.or.th/assets/portals/1/files/Multi_sided_Platform.pdf (last visited Sept. 30, 2026).
[85] Trade Competition Comm’n of Thai., TCCT to Revise Guidelines on Digital Platforms, Wholesale-Retail Trade, and On-Demand Delivery Services, Press Release No. 12/2569 (July 31, 2026), https://www.tcct.or.th/assets/portals/1/files/Press_12_69.pdf.
[86] Trade Competition Comm’n of Thai., TCCT Establishes Two Subcommittees to Intensify Oversight of Digital Platforms and Set Modern Trade Rules, Press Release No. 9/2569 (July 7, 2026), https://www.tcct.or.th/assets/portals/1/files/Press_09_2569.pdf.
[87] Joseph O’Connor, Bhumjaithai Led Government Targets Online Platforms with New Digital Platform Bill with Stronger Powers, Thai Exam’r (July 20, 2026), https://www.thaiexaminer.com/thai-news-foreigners/2026/07/20/bhumjaithai-led-government-targets-online-platforms-with-new-digital-platform-bill-with-stronger-powers.
[88] The Nation Editorial Team, Digital Minister Slows Platform Bill and Rethinks Fees Charged to Businesses, Nation Thai. (Aug. 14, 2026), https://www.nationthailand.com/news/policy/40069775.
[89] Ministry of Indus. & Trade of Viet., Draft Decree on Administrative Penalties for Violations in the Field of Competition (Sept. 22, 2026), https://moit.gov.vn/du-thao-van-ban/du-thao-nghi-dinh-quy-dinh-ve-xu-phat-vi-pham-hanh-chinh-trong-linh-vuc-canh-tranh.html.
[90] Indonesia Antitrust Regulator Proposes Digital Markets Law over AI, Platform Risks, MLex (May 27, 2026), https://www.mlex.com/mlex/articles/2482120/indonesia-antitrust-regulator-proposes-digital-markets-law-over-ai-platform-risks.
[91] See also Göhsl, supra note 64 (arguing that uncertainty about the DMA’s scope and its ability to address emerging gatekeepers leaves a vital role for competition law, particularly through more forward-looking enforcement).
[92] DMA Review Staff Working Document, supra note 8, § 4.2.3. See also DMA Review Report, supra note 25, at 10.
[93] European Comm’n, Commission Imposes Interim Measures on Meta to Preserve Free Access to WhatsApp for Rival AI Assistants (June 9, 2026), https://ec.europa.eu/commission/presscorner/detail/en/ip_26_1276 [hereinafter Meta WhatsApp Interim Measures]; European Comm’n, Commission Opens Investigation into Possible Anticompetitive Conduct by Google in the Use of Online Content for AI Purposes (Dec. 9, 2025), https://ec.europa.eu/commission/presscorner/detail/en/ip_25_2964.
[94] See, e.g., Autorité de la Concurrence, Opinion on the Competitive Functioning of the AI Agents Sector, Opinion No. 26-A-05, ¶ 354 (July 17, 2026), https://www.autoritedelaconcurrence.fr/en/opinion/competitive-functioning-artificial-intelligence-ai-agents-sector.
[95] Meta WhatsApp Interim Measures, supra note 93; Autorità Garante della Concorrenza e del Mercato, Meta AI: The Italian Competition Authority Orders Meta to Suspend the Terms Excluding Competing AI Chatbots from WhatsApp (Dec. 24, 2025), https://en.agcm.it/en/media/press-releases/2025/12/A576.