Matt Sheehan and Scott Singer
The Asia Program in Washington studies disruptive security, governance, and technological risks that threaten peace, growth, and opportunity in the Asia-Pacific region, including a focus on China, Japan, and the Korean peninsula.Learn More
The Technology and International Affairs Program develops insights to address the governance challenges and large-scale risks of new technologies. Our experts identify actionable best practices and incentives for industry and government leaders on artificial intelligence, cyber threats, cloud security, countering influence operations, reducing the risk of biotechnologies, and ensuring global digital inclusion.Learn More
China’s most influential AI standards body released a comprehensive articulation of how technical experts and policy advisers in China understand AI risks and how to mitigate them.
The AI Safety Governance Framework 2.0,1 released in September, builds on an earlier version of the framework released a year prior. Alongside the Chinese Communist Party’s (CCP) unwavering focus on “information content risks” from AI, Framework 2.0 responds to the advances of AI over the past year, such as the global proliferation of open-source models and the advent of reasoning models. It represents a significant evolution in risks covered, including those tied to labor market impacts and chemical, biological, radiological, and nuclear (CBRN) weapon misuse. And it introduces more sophisticated risk mitigation measures, establishing a rubric to categorize and grade AI risks that sector-specific regulators should adapt to their domain.
The framework is not a binding regulatory document. But it offers a useful datapoint on how China’s AI policy community is thinking about AI risks. It could also preview what technical AI standards—and possibly regulations—are around the corner. Given China’s massive footprint in AI development, the impact of those standards will ripple out across the world, affecting the trajectory of the technology itself.
Who’s Behind the Framework?
No comments:
Post a Comment