Evolutionary AI in Practice: Japan’s Bottom-Up Innovation Model and the Co-Evolution of Industrial and Competition Policy
Abstract
This Article argues that Japan’s trajectory in artificial intelligence is best understood as an evolutionary process in which bottom-up deployment of AI across industrial and service domains co-evolves with a policy framework that integrates restrained industrial support and anticipatory competition enforcement. In contrast to the centralized foundation-model strategies pursued in the United States and China and the regulatory-led framework emerging in the European Union, Japan’s approach highlights how sustained deployment, domain specialization, and incremental learning can reshape both industrial structure and the focus of competition policy. Drawing on developments across robotics, manufacturing, mobility, and enterprise software, the Article introduces the concept of the deployment layer to explain how value creation and competitive bottlenecks increasingly arise at the interfaces through which AI systems are operationalized rather than within models alone. This perspective reframes industrial policy debates by suggesting that compute access, experimental deployment, and organizational learning may matter more than large-scale frontier-model subsidies, while competition policy must increasingly address control over data, infrastructure, and integration pathways. The Article concludes that Japan’s experience illustrates an alternative governance model for AI in which industrial policy and competition enforcement evolve together through iterative market learning, offering a framework for understanding AI governance as a process of policy co-evolution rather than discrete regulatory intervention.
Read the full piece at SSRN.