Amazon warning triggered US crackdown on Anthropic AI models: Reports
Amazon warned US officials that Anthropic's Fable 5 could be prompted into cyberattack-useful output, reportedly triggering a foreign-access cutoff.
Intelligence analysis by GPT-5.4 Mini
A Wall Street Journal report says Amazon CEO Andy Jassy's warning helped prompt the Trump administration to suspend foreign access to Anthropic's Fable 5 and Mythos 5 models. The episode also lifted decentralized AI tokens after the market read it as evidence that access to US-based AI systems can be switched off quickly.
A cloud company said a smart computer tool might be tricked into helping bad actors, so the government locked it from foreign users. Crypto traders then bought coins tied to decentralized AI, like switching to a backup generator when the main power line looks risky.
Analysis
What happened
According to the report cited by Cointelegraph, Amazon CEO Andy Jassy contacted senior US officials after Amazon researchers found a way to prompt Anthropic’s Fable 5 model into giving information that could support cyberattacks. The warning, together with concerns from at least five other firms, reportedly led to a rapid White House response and a request to Anthropic CEO Dario Amodei to voluntarily pull the model.
The administration then issued an export control that cut off foreign access to Anthropic’s newest models, Fable 5 and Mythos 5, on Friday night. David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, said the administration acted reluctantly and wanted Anthropic to fix the jailbreak issue so access could be restored.
Anthropic’s response
Anthropic pushed back on the government’s interpretation. In a Friday blog post, the company said it believed the directive came from a misunderstanding of a “non-universal jailbreak” described in an unnamed report. Anthropic also said it is working to restore access for users.
Amazon, for its part, did not confirm whether it spoke to government officials about the models. A spokesperson said governments sometimes seek its counsel on security risks, but the company does not share details of those discussions.
Crypto market reaction
The story spilled into crypto because traders treated the crackdown as proof that centralized AI access can be restricted quickly. Bittensor rose 23.9% over 24 hours, Venice Token gained 16%, and Near Protocol climbed 6.2%, according to the article. That move reflected a broader trade around decentralized AI infrastructure after the US action.
Key points
- Amazon reportedly warned US officials after researchers found a way to prompt Anthropic's Fable 5 into producing cyberattack-useful information.
- The Trump administration then cut off foreign access to Anthropic's Fable 5 and Mythos 5 models.
- David Sacks said the administration acted reluctantly and wanted Anthropic to fix the jailbreak issue.
- Anthropic said the directive stemmed from a misunderstanding about a non-universal jailbreak.
- Bittensor, Venice Token, and Near Protocol rose after the crackdown as traders favored decentralized AI narratives.
If Anthropic fixes the safety issue, the article says access could be restored and Fable 5 could return to general release. That would reduce uncertainty for users and show that the model can be corrected without a long shutdown. Decentralized AI projects could also keep benefiting from renewed interest in systems that are harder to switch off.
If regulators stay unconvinced, foreign access could remain restricted for longer and Anthropic’s rollout could stay disrupted. The article also suggests the episode may reinforce the idea that centralized AI platforms are vulnerable to sudden policy shocks, which could keep crypto traders rotating into speculative AI tokens rather than fundamentals.
