Anthropic Adds Security Measure to Regain Trump Administration's Favor

The Trump administration lifted export controls on Anthropic's Claude Fable 5 AI model after the company agreed to extend an existing guardrail to prevent users from trying to access certain restricted capabilities, according to two people familiar with the matter. The safeguard means any users trying to unlock those capabilities will be notified that their request is blocked and will have their query processed by the less-advanced Opus 4.8 AI model, the people say. Before Anthropic cut off access to Fable 5, user requests related to sensitive cybersecurity and biology capabilities were supposed to be processed by Opus 4.8.

The new safeguard, the people say, will extend this guardrail to requests related to a specific behavior identified in a paper by Amazon. According to an analysis published by Katie Moussouris, founder and CEO of Luta Security, after reading the Amazon paper, users were able to get around a restriction on Fable 5 by asking the model to fix code, rather than identify security issues in it. While cybersecurity experts generally don't find this behavior troubling, the administration learning about it led to the showdown with Anthropic and the imposition of export controls, which as a practical matter took the model offline.

The addition offers new detail to Commerce secretary Howard Lutnick's letter announcing the removal of restrictions on Anthropic's Fable 5 and Mythos 5 AI models. "Among other things, Anthropic has agreed to proactively detect and address security risks posed by the models," wrote Lutnick, who led the effort to bring the models back online. WIRED first obtained the letter and shared its details on Tuesday night.

The Commerce Department also ultimately cleared Fable 5 for release after researchers at its Center for AI Standards and Innovation decided the safeguards on the model were sufficiently robust for now, the people added. Still, while Anthropic has resolved its impasse with the Commerce Department, defense secretary Pete Hegseth has told advisers there is no clear path to lift his February 28 order designating the company a supply chain risk, according to a person briefed on the matter. So while some of Anthropic's challenges with the administration are less pressing, they're not entirely over.

The Trump administration's decision to lift export controls on Anthropic's AI model has significant implications for the broader AI industry. By requiring Anthropic to extend its guardrail to prevent access to restricted capabilities, the administration is setting a precedent for AI developers to prioritize security and safety. This move could lead to increased scrutiny of AI models and their potential misuse, potentially slowing down the development and deployment of AI technologies.