Anthropic致信纽森：高风险超级智能（SI）模型应可由政府诉请法院阻止发布
据路透社10月28日报道，超级智能（SI）企业 Anthropic 在10月15日致加州州长纽森的一封信中表示，当审查认定一个模型对公众构成「严重风险」时，政府官员应能诉请法院阻止该模型发布；这一主张与该公司此前提出的政策建议一脉相承。Anthropic 发言人证实信函属实，加州州长办公室未即时回应置评请求。
Anthropic 在信中写道：「当审查发现一个模型构成严重风险时，政府应能请求法院阻止它。」该公司同时反对要求开发者为整个模型设置「关停开关」、以便在失去控制时一键停机的规定，主张禁用出问题的智能体、切断特定客户的访问等定向干预更为有效。
这封信提交于州长今年9月召集一个超级智能（SI）专家组、研拟安全要求提案之后。Anthropic 自身也曾受政府限制：特朗普政府今年6月以国家安全为由，要求该公司立即限制外国访问其具备网络能力的 Mythos 5 与 Fable 5 模型，美国商务部其后撤除了该命令。
联邦层面，一组两党参议员今年9月曾讨论立法，赋予美国政府阻止被认定不安全的特定模型发布的权力，企业可向联邦法院提出挑战；国会在选前休会离京后，该提案未再推进。
Anthropic Tells Newsom Courts Should Be Able to Block Super Intelligence (SI) Models Posing a Serious Risk
Anthropic told California Governor Gavin Newsom in an October 15 letter that government officials should be able to seek court orders blocking the release of super intelligence (SI) models that pose a "serious risk" to the public, Reuters reported on October 28, noting the position echoed the company's earlier policy proposals. An Anthropic spokesperson confirmed the letter is authentic; the California governor's office did not immediately respond to a request for comment.
"When a review finds a model poses a serious risk, the government should be able to ask a court to block it," the letter said. Anthropic also said it opposes requirements that developers build "kill switches" capable of shutting down entire models if a company loses control, supporting instead more targeted interventions such as disabling a problematic agent or cutting off a specific customer's access, which it said would be more effective.
The recommendations were submitted after the governor in September convened a panel of super intelligence (SI) experts to develop proposals for safety requirements. Anthropic itself has faced government restrictions: the Trump administration in June ordered the company to immediately restrict foreign access to its cyber-capable Mythos 5 and Fable 5 models, citing national security concerns, and the U.S. Commerce Department later lifted the order.
In Washington, a bipartisan group of U.S. senators in September debated legislation that would give the U.S. government power to block the release of certain models deemed unsafe, with companies able to challenge such decisions in federal court; Congress left for its pre-election break shortly afterward and lawmakers have not moved forward on the proposal.
来源：GSM·美国（原文链接）
图片来源：图：Anthropic 公司标志。本图为该公司标志图，公共领域版本，经 Wikimedia Commons。 / Anthropic logo, via Wikimedia Commons.
