特朗普宣布'超级智能'协议
特朗普推动的AI安全协议，六家科技巨头签署，靠同行压力而非法律约束
特朗普在Truth Social上发布'超级智能'协议，谷歌、OpenAI、Anthropic等六家科技公司签署。该协议要求各公司建立强大的内部AI安全控制系统，组建内部团队运行，聘请外部审计师检查工作，并成立董事会委员会监督。扎克伯格表示签署方同意董事会将独立审查审计报告，使公司董事成为最后一道审查防线。
President Trump just announced ‘Super Intelligence’ Accord, with his latest post on TruthSocial
Google, OpenAI, Anthropic, Meta, xAI and Nvidia executives signed a White House AI safety accord
Trump posted the one-page document on Truth Social, and it asks each participating company to adopt 4 safeguards for the technology. Yahoo Finance
Each company commits to build robust internal controls for frontier AI safety, staff internal teams to run them, hire outside auditors to check that work and create a board committee to oversee it.
Mark Zuckerberg said the signatories agreed their boards of directors would independently review the auditors' reports, which makes corporate directors the last line of review.
Trump's enforcement mechanism is peer pressure, since he described the accord as "morally binding" and said it would keep companies in line through "self-policing and group policing."
The document itself says it might later make sense to turn these steps into laws or regulations, but House members aren't expected back in Washington until after the November midterms.
