For the makers and artists building the next generation of tools, the landscape has just shifted from open experimentation to high-stakes compliance. On Friday, barely a week after Anthropic released its new Fable 5 and Mythos 5 models to the public, the US government moved to impose strict export controls on them. These restrictions barred access for foreign nationals, including non-US citizens working at Anthropic. Consequently, Anthropic took both models offline for everyone, citing an inability to reasonably enforce the order without shutting them down entirely.
The situation remains fluid. As of Tuesday, Fable is still inaccessible. Users attempting to boot up Claude are met with a notification stating that “Fable 5 is currently unavailable.” While the immediate future of the model’s return is uncertain, the implications for the tech industry and the nation’s regulatory framework are profound.
The irony of the safety argument
There is a significant irony at the heart of this unfolding drama. Anthropic has spent years arguing that AI models could soon become so powerful they pose a genuine danger to society, urging the government to implement serious regulation sooner rather than later. Now, with the administration taking a hardline stance, the company finds itself at odds with the very policies it advocated for.
This event will be watched closely by global powers, particularly in China, to determine if the US approach evolves into a robust safety framework or devolves into a political weapon used to punish companies that do not immediately align with the administration’s demands.
The timeline of the shutdown
To understand the chaos, one must look at the sequence of events. Anthropic initially released Mythos 5 and the safeguarded Fable 5, which was designed to be a watered-down version of the more dangerous underlying model. While the company had previously warned that Mythos could act as a cyber-weapon, Fable was intended to be safe enough for public use.
However, the release was not without controversy. Independent researchers and red teamers found the safety guardrails on Fable to be so restrictive that they hindered legitimate research, often downgrading interactions to older models when users pushed boundaries.
The turning point occurred when Amazon researchers uncovered a potential jailbreak for the system. Reports indicate that Amazon CEO Andy Jassy became concerned and contacted members of the Trump administration, including Treasury Secretary Scott Bessent, to discuss the risk. The response was swift and severe: the administration issued a 90-minute ultimatum to Anthropic to shut down the models.
Anthropic scrambled to respond, contacting the administration within 15 minutes to seek clarification on the specific vulnerability. However, the window closed before a resolution could be reached. The administration proceeded with export controls, forbidding any foreign nationals from accessing the models, regardless of their employment status. This meant that even US-based employees who were not citizens could no longer use the systems.
What this means for the industry
The fallout extends beyond a single product launch. It highlights a friction point between corporate safety protocols and government enforcement capabilities. Anthropic had spent considerable time hyping the dangers of these models to justify their release, yet the government’s reaction suggests that the perceived risk has escalated beyond the company’s initial assessments.
As Hayden Field, a senior AI reporter, noted, the weekend scramble has left the regulatory path unclear. The administration’s willingness to act so quickly, and the subsequent inability to manage the restrictions without a total shutdown, signals a new era of volatility for AI developers.
Key takeaways
- Anthropic has taken Fable 5 and Mythos 5 offline in response to a 90-minute ultimatum from the Trump administration following reports of a potential jailbreak.
- New export controls prohibit foreign nationals, including non-citizen employees, from accessing the models, creating a compliance crisis for the company.
- The incident underscores the growing tension between AI companies advocating for regulation and the government’s increasingly aggressive approach to AI safety.
- Industry observers are now watching to see if this event marks the beginning of a formal safety framework or merely a tool for political leverage against tech firms.
