安远AI：2024中国人工智能安全全景报告（英文版）（114页）.pdf

1、The State of AI Safety in ChinaSpring 2024 ReportPublished May 14,20242Executive Summary(1)The relevance and quality of Chinese technical research for frontier AI safety has increased substantially,with growing work on frontier issues such as LLM unlearning,misuse risks of AI in biology and chemistr

2、y,and evaluating power-seeking and self-awareness risks of LLMs.There have been nearly 15 Chinese technical papers on frontier AI safety per month on average over the past 6 months.The report identifies 11 key research groups who have written a substantial portion of these papers.Chinas decision to

3、sign the Bletchley Declaration,issue a joint statement on AI governance with France,and pursue an intergovernmental AI dialogue with the US indicates a growing convergence of views on AI safety among major powers compared to early 2023.Since 2022,8 Track 1.5 or 2 dialogues focused on AI have taken p

4、lace between China and Western countries,with 2 focused on frontier AI safety and governance.Executive Summary3Executive Summary(II)Chinese national policy and leadership show growing interest in developing large models while balancing risk prevention.Unofficial expert drafts of Chinas forthcoming n

5、ational AI law contain provisions on AI safety,such as specialized oversight for foundation models and stipulating value alignment of AGI.Local governments in Chinas 3 biggest AI hubs have issued policies on AGI or large models,primarily aimed at accelerating development while also including provisi

6、ons on topics such as international cooperation,ethics,and testing and evaluation.Several influential industry associations established projects or committees to research AI safety and security problems,but their focus is primarily on content and data security rather than frontier AI safety.In recen