Agile and iterative governance: China’s regulatory response to AI

Generative AI has been a buzzword in various sectors, advancing at an exponential rate and profoundly transforming the way we communicate and innovate. However, its potential benefits come with compelling ethical and legal risks, necessitating proper guardrails steering AI in beneficial directions. Amid the global race to AI regulation, China has exhibited strong and open ambition in shaping the emerging global AI order. In response to challenges posed by AI, China has not only implemented agile administrative intervention in policy support, central coordination, and investment, but also adopted an AI governance framework that characterized by ‘complexity,’ ‘agility,’ ‘stability,’ and ‘flexibility.’ Nevertheless, while an agile, sector-specific approach to AI governance may yield short-term benefits, it raises long-term concerns, including opaque decision-making, weak enforcement, fragmented oversight, and inadequate protection of fundamental rights. In particular, governance fragmentation marked by overlapping regulatory bodies and layered rulemaking risks producing piecemeal and outdated regulations that struggle to keep pace with rapid technological change. Current interventions often prioritize systemic stability over ethical clarity and robust supervision, while strategic ambiguity further complicates the implementation of AI ethics and hinders effective oversight, whether internal or administrative.

Instead of calling for an omnibus AI law that applies a uniform package of rules, Chinese regulators chose to adapt horizontal elements into vertical regulations through a set of bureaucratic know-how and iterative regulatory tools. Through a comparative legal analysis, this paper finds that comprehending the intricacies of China’s AI regulatory approach is vital not only for projecting its future technological progression but also for understanding its impact on international tech competition. Differences in diverse AI governance may offer valuable insights while commonalities in AI governance values and principles hold promises for global cooperation in responsible AI governance. Moreover, it is plausible to expect that China will reconcile existing horizontal regulatory tools in a horizontal legislative package, while the EU AI Act provides valuable practical implications for the transition from vertical AI-related regulations to a horizontal Chinese AI Law, and the decentralized regulatory approach adopted by the U.S. serves as a useful reference for multi-stakeholder and multi-level cooperation. In addition, EU’s rights-driven framework and US’s market-driven model may serve as critical benchmarks, influencing China’s state-driven approach to harmonizing its legislative strategies for a responsive AI regulation.