AI regulation is the body of government requirements governing how artificial-intelligence systems are developed, deployed, disclosed, and used. It seeks to balance innovation with responsible and ethical practice, especially around privacy, transparency, bias, and accountability. section 10.28
Regulation operates within a wider governance system. Government rules may be informed by industry, academia, and civil society, while corporate self-governance can complement formal requirements. Effective oversight must address both technical risks and social consequences as generative AI becomes more integrated, efficient, personalized, and agentic. section 10.0 section 10.35
Major regulatory approaches differ by jurisdiction:
- The European Union uses a risk-based, horizontal framework intended to cover applications of a technology broadly. section 10.31 section 10.35
- The United States has not pursued comparably large-scale regulation and has emphasized innovation and development through an AI Action Plan. section 10.35
- China uses an iterative, vertical approach that targets particular applications and can be updated as concerns or gaps emerge. section 10.31
China’s targeted rules
By 2023, China had introduced regulations addressing recommendation algorithms, deep synthesis such as deepfakes, and generative AI. These rules established requirements for system deployment and disclosures to government and the public. Recommendation rules also addressed algorithmic transparency, user control over data, information management, and manual intervention. section 10.31
