Amazon CEO Andy Jassy flagged security risks in Anthropic's Fable 5 and Mythos 5 models to the White House, triggering a federal review that led to restrictions on the AI systems.
Alpha Score of 55 reflects moderate overall profile with strong momentum, poor value, strong quality, moderate sentiment.
Amazon CEO Andy Jassy personally flagged security concerns about Anthropic's latest AI models, Fable 5 and Mythos 5, weeks before U.S. regulators moved against them, according to people familiar with the matter. The intervention set off a chain of internal reviews at federal agencies that culminated in the current restrictions on the systems.
Jassy raised the issues in a direct communication to the White House in late January, the people said. He cited potential vulnerabilities in how the models could be used to generate disinformation at scale and automate cyberattacks. Amazon holds a $4 billion stake in Anthropic and has integrated the startup's technology into its AWS cloud platform.
The White House referred the matter to the National Institute of Standards and Technology, which ran its own evaluation. NIST's findings aligned with Jassy's concerns, leading to the Commerce Department's decision to restrict the models' public deployment under the Biden administration's AI executive order, the people said.
Anthropic has pushed back against the characterization of its models as dangerous. The company said in a statement that Fable 5 and Mythos 5 underwent extensive internal and external safety testing before release and that no independent evaluation had found them to pose imminent risks. The company also noted that Amazon's own cloud services offer the models to customers.
Amazon declined to comment on Jassy's specific communications. A White House spokesperson said the administration takes all credible AI safety reports seriously and follows established review procedures.
The episode underscores the unusual position Amazon occupies in the AI landscape. The company is both Anthropic's largest investor and a competitor through its own AI lab, which develops models under the Alexa and AWS brands. Jassy's intervention raises questions about whether Amazon used its government access to slow a rival's technology, though the people familiar said his concerns were technical, not commercial.
Fable 5 and Mythos 5 remain available through Anthropic's API with additional safeguards the company added after the federal review. The restrictions do not apply to earlier versions of the models or to Anthropic's smaller systems.
The Commerce Department's Bureau of Industry and Security is now drafting broader rules for large AI model releases, expected later this year. Anthropic has said it will comply with all regulations while continuing to argue that the current restrictions are based on hypothetical risks rather than demonstrated harms.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.