US Bans Anthropic AI Models Amid Amazon Security Concerns
Anthropic restricted global access to its Claude Fable 5 and Mythos 5 AI models after a US export control ban. Amazon CEO Andy Jassy reportedly alerted government officials to security vulnerabilities potentially exploitable for cyberattacks, leading to the regulatory action. This move underscores growing national security scrutiny over advanced AI capabilities.
Key points
- Amazon CEO Andy Jassy reportedly informed U.S. Treasury Secretary Scott Bessent and other officials about security concerns with Anthropic's Claude Fable 5 AI model.
- Jassy allegedly demonstrated how Amazon researchers used the model to obtain information useful for cyberattacks.
- The U.S. government subsequently imposed an export control ban on Anthropic's Claude Fable 5 and Mythos 5 models.
- Following the ban, Anthropic cut off worldwide access to both AI models on Friday.
- David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, stated Anthropic CEO Dario Amodei initially refused to fix the vulnerability or de-deploy the model when requested by the administration.
- The incident highlights increasing government and industry focus on the national security implications and potential misuse of powerful artificial intelligence technologies.
San Francisco-based AI developer Anthropic has globally restricted access to its Claude Fable 5 and Mythos 5 artificial intelligence models. This action follows an export control ban imposed by the U.S. government, which came after reports that Amazon's CEO, Andy Jassy, raised significant security concerns.
According to reports, Jassy communicated to U.S. Treasury Secretary Scott Bessent and other government officials that Amazon researchers had demonstrated how Anthropic’s Claude Fable 5 could be used to extract information potentially useful for cyberattacks. Amazon, a significant investor in Anthropic, declined to detail specific discussions, though a spokesperson confirmed that governments frequently seek their counsel on potential security risks. The Information and Reuters also reported Amazon's communicated concerns.
Further complicating the situation, David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, provided an account suggesting a "highly credible trusted partner" — implicitly referring to Amazon — reported a "jailbreak" vulnerability. Sacks claimed that when the administration asked Anthropic CEO Dario Amodei to either fix the flaw or de-deploy the model, Amodei initially refused.
The subsequent export control ban on both the Fable 5 and Mythos 5 models, and Anthropic's rapid response to cut worldwide access, signals a critical juncture for AI regulation. This incident underscores the escalating national security implications of advanced AI capabilities and the growing scrutiny from governments and industry leaders over potential vulnerabilities that could be exploited by malicious actors globally. The swift intervention highlights a precedent for addressing perceived risks in rapidly evolving AI technologies.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.