Anthropic executives are meeting with the Trump administration on Monday, according to multiple reports, just days after an order forced the company to disable its Fable 5 and Mythos 5 models for public use—the latest escalation in the conflict between the U.S. government and the AI giant reportedly planning an IPO later this year.
Days earlier, the Trump administration ordered Anthropic to ban foreign nationals from using its Claude Mythos models.
NurPhoto via Getty Images
Executives at Anthropic are planning to meet with members of the Commerce Department on Monday, multiple outlets reported, although the proceedings are reportedly confidential and it is unclear which Anthropic staffers will take part in the discussions.
The meeting comes only three days after the Trump administration issued a directive ordering Anthropic to ban foreign nationals—including foreign nationals who work for the company—from accessing its newest Claude models.
Claude Fable 5, the first publicly available version of the company’s purportedly powerful and possibly dangerous Mythos model, was only released last Tuesday.
The company said the export control directive forced it to pull access to Fable 5 for all customers, both inside and outside the U.S.
Members of Anthropic’s technical staff have been continuously meeting with Trump administration officials since the export control on Friday, Reuters reported.
Neither Anthropic nor the Commerce Department immediately returned a request for comment from Forbes.
Fable 5 was released to the public on Tuesday with what Anthropic called “strong safeguards to prevent misuse.” These essentially prevented users from entering prompts about certain topics, like cybersecurity needed to develop advanced hacking techniques and biology needed to develop bioweapons. On Friday, Anthropic said in a blog post it believed the U.S. government was shown a way to “jailbreak” the Fable model, or bypass its security safeguards. “We suspect that perfect jailbreak resistance is not currently possible for any model provider,” the company admitted in the blog post, and said instead it developed a “defense in depth” strategy to detect jailbreak attempts or make universal jailbreaks too expensive to produce.
The abrupt export control order and the subsequent takedown of Fable was only the latest move in the ongoing conflict between the Trump administration and Anthropic. In March, the Pentagon designated the company a supply chain risk—an unusual move which effectively barred the military from using its AI models for Defense Department work. Anthropic has challenged the designation in court, but the Defense Department has remained adamant about retaining the risk designation. “Three months ago, @DeptofWar kicked @AnthropicAI out of our building—forever,” Defense Secretary Pete Hegseth said in a post on X on Friday, hours after the export control was issued. “Every passing day proves why that was the right move.”
ForbesAnthropic’s Fable 5 Safeguards Were Always A ‘Judgement Call’
ForbesOpenAI, Nvidia, Alphabet And More Sign AI Deal With Pentagon For Classified Military UseBy Ty RoushForbesAnthropic CEO: ‘We Don’t Know Exactly How’ Claude AI Was Used In Iran School StrikeBy Antonio Pequeño IV
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。