美国政府近期禁止外国籍人士使用Anthropic最新推出的模型Mythos和Fable。此举引发了矽谷和欧洲的担忧，部分科技界人士将此决策归咎于Anthropic对AI风险的过度渲染，AI先驱Yann LeCun更直言这是其「荒谬的恐慌煽动」付出的代价。然而，Anthropic执行长Dario Amodei坚称Mythos在网路安全和国家关键基础设施方面确实存在实质且不可忽视的风险。
除了内部的安全警告，Anthropic与美国政府的关系也日益紧张，先前五角大厦曾将其列为国家安全供应链风险，双方目前正为此进行诉讼。尽管YouGov民调显示多数美国民众支持加强AI监管，但独立政策研究员指出，一个自称支持创新且推动晶片出口的政府，却禁止了国内最先进的AI模型，这种矛盾的政策反应令人感到困惑。
According to a Financial Times analysis, Anthropic has warned about the dangers of advanced AI far more frequently than its rival OpenAI in 2026. The research shows that five out of every 1,000 words in Anthropic's official statements and social media posts were related to risk, regulation, or restrictions, which is eight times higher than OpenAI's rate. This warning rhetoric has sparked controversy, with critics accusing Anthropic of helping to trigger a US ban on foreign access.
The US government recently barred foreign nationals from accessing Anthropic's latest models, Mythos and Fable. This decision has alarmed Silicon Valley and Europe, with some technologists blaming the export ban on Anthropic's repeated risk warnings, and AI pioneer Yann LeCun calling it a consequence of Amodei's "ridiculous fear-mongering." However, Anthropic CEO Dario Amodei maintained that Mythos poses real cybersecurity and infrastructure risks.
Beyond its internal safety warnings, Anthropic's relationship with the US government has grown tense, following the Pentagon designating it as a supply-chain risk to national security in February, which led to litigation. Although YouGov polling shows that the majority of the US public supports effective AI regulation, independent policy researchers argue that it does not inspire confidence when a pro-innovation administration bans its own country's most advanced model.