Archive — US — techcrunch
Amazon CEO reportedly raised Anthropic model concerns before governmentcrackdown
Amazon CEO Andy Jassy raised concerns over Anthropic's AI models, leading to a government ban on access to Claude Fable 5 and Mythos 5 due to potential cybersecurity risks. An Amazon spokesperson confirmed the company's engagement with government on security, and Anthropic claimed that the concerned capabilities are already present in publicly available models.
Key points
- Andy Jassy informed U.S. officials about security risks posed by Anthropic's AI models.
- This led to a government export control ban on Claude Fable 5 and Mythos 5.
- Amazon, a significant investor in Anthropic, was cited as having communicated these risks.
- David Sacks mentioned a credible partner raised issues about security, asking Anthropic to fix or remove risky capabilities.
- Anthropic stated that the features causing concern are already available in other models.
Why it matters
- Concerns about AI safety and cybersecurity are increasingly prominent within AI governance discussions as of 2025.
- Similar export controls have emerged in other countries focusing on AI technologies, reflecting a global trend.
- The incident highlights the ongoing conflict between rapid AI development and regulatory oversight in the tech industry.
The interaction between Amazon and government officials underscores the tension between AI innovation and security, mirroring broader industry trends towards stricter regulations and safety measures in AI deployment.