OpenAI 聘請曾警告 AI 可能導致人類滅絕的數學家
菲爾茲獎得主 jacob tsimerman 加入了 openai 的安全團隊，儘管他此前合著了一篇關於 AI 導致人類滅絕風險的論文，凸顯了黑盒模型中安全保證的激烈爭論。
fields medalist jacob tsimerman joined openai's safety team to address the lack of theoretical guarantees in empirical ai systems. he previously co-authored a paper analyzing 'omnicide events' where ai contributes to human extinction, arguing society must assess risks honestly rather than panic. tsimerman predicts ai will soon outperform humans in mathematics, though demis hassabis notes current models like openai's o1 haven't solved millennium prize problems yet. this move underscores a growing trend of elite mathematicians entering ai labs to formalize safety.