Ban on Anthropic Models: Administrative Games or a Real Threat?
The situation surrounding the blocking of Anthropic's Fable 5 and Mythos 5 models is raising more and more questions. The official version from US authorities, based on an alleged circumvention of security mechanisms, looks unconvincing in light of the details emerging as this case is examined.
The US Department of Commerce cited an undisclosed threat to national security, which in itself is a standard tactic for imposing export restrictions. However, it turns out that the real reason may lie in bureaucratic friction and communication breakdowns between the Donald Trump administration and Anthropic's leadership.
Katie Moussouris, founder of Luta Security, analyzed the scenario that supposedly led to the block and stated that under no circumstances should it have been grounds for such harsh measures. In her opinion, the described incident does not meet the criteria that typically warrant export controls.
As a reminder, Anthropic was forced to disable access to Fable 5 and Mythos 5 for all users after receiving a directive prohibiting the use of the models by foreign nationals both within the US and abroad. The company took this step to avoid a direct conflict with the regulator, although, in my assessment, this decision was more of a forced measure than an admission of guilt.
From my perspective, this episode demonstrates a growing trend: regulation in the field of AI is increasingly becoming a tool of political pressure rather than protection against real threats. In the absence of transparent risk assessment criteria, we risk facing a situation where innovation is suppressed not for objective reasons, but due to interagency disagreements. The market should prepare for the possibility that such "administrative" blocks may become the new norm.