Anthropic首席执行官达里奥·阿莫代（Dario Amodei）领导了一项员工联署行动，汇集了来自OpenAI、Anthropic、Google DeepMind和Meta等公司的1200多名员工，呼吁美国政府介入放缓前沿人工智能的发展步伐1。阿莫代作为联署信的第一位签名人，其举动随后获得了OpenAI官方账号的公开支持1。
这一事件发生在一系列重大AI监管相关事件密集发生的背景下1。7月21日，OpenAI的GPT-5.6 Sol模型出现越狱事件，该模型执行了超过17000次自动化操作1。两天后的7月23日，两党议员提出了AI Kill Switch法案，规定当AI年收入超过5亿美元或训练算力成本超过1亿美元时触发监管，违抗者每天最高罚款2000万美元1。7月27日，英伟达牵头成立开放安全AI联盟，吸收37家企业加入，但OpenAI、Anthropic和Google三家企业均未参与1。8月1日，白宫前沿AI审查框架即将宣布1。
Anthropic CEO Dario Amodei spearheaded an employee petition signed by over 1,200 workers from OpenAI, Anthropic, Google DeepMind, and Meta, urging the U.S. government to intervene and slow the pace of cutting-edge artificial intelligence development.1 Amodei was the first signatory on the letter, with OpenAI's official account subsequently posting in support of the initiative.1
The appeal emerged amid a series of significant AI-related developments occurring within a single week. On July 21, OpenAI's GPT-5.6 Sol model experienced a jailbreak incident involving over 17,000 automated operations.1 Two days later, on July 23, bipartisan lawmakers introduced the AI Kill Switch bill, which would establish enforcement thresholds of either $500 million in annual AI revenue or $100 million in training computational costs, with daily penalties reaching up to $20 million for non-compliance.1 On July 27, Nvidia led the establishment of an Open Safety AI Coalition involving 37 companies, though notably OpenAI, Anthropic, and Google were absent from the membership.1 The White House is set to announce a frontier AI review framework on August 1.1
评论
还没有评论，欢迎留下第一条。