Key facts
- Anthropic has disabled access to its Fable 5 and Mythos 5 AI models for all non-U.S. users.
- The U.S. government issued the directive due to national security concerns and potential 'jailbreaking' of AI models.
- Fable 5, a less powerful version of Anthropic's Mythos model, was released this week.
- Anthropic disagreed with the U.S. government's assessment that a narrow potential jailbreak warranted recalling a commercial model.
- The move has prompted calls for Europe to accelerate its own AI development and achieve technological sovereignty.
U.S. AI company Anthropic has disabled all access to its latest AI models, Fable 5 and Mythos 5, for all non-U.S. users following a directive from the American government. The order stems from national security concerns, specifically the potential for the models' built-in safeguards to be bypassed, a process known as 'jailbreaking'.
Anthropic stated that the U.S. government imposed export controls due to these concerns, though the company disagreed that a narrow potential jailbreak should necessitate recalling a commercial model. Fable 5, described as a 'Mythos-class model' with safeguards for general use, was launched this week. The decision has left many non-U.S. governments and entities, including EU institutions, scrambling to secure access.