Scoop: Trump admin blocks foreign access to Anthropic's most powerful AI
TL;DR
The US government is moving to block foreign access to Anthropic's Fable 5 and Mythos 5 models. The restriction would cover foreign governments, companies, individuals and foreign nationals inside the US. Anthropic responded by disabling the models for all customers for now, saying this is necessary to comply with the export-control order. Access to other Anthropic models is not affected.
Nauti's Take
This shows how thin the line between AI safety policy and industrial policy has become. Anthropic has long pushed for stricter oversight, but now faces a rule it describes as unfair and poorly grounded.
If governments treat frontier models like cyber weapons, they need solid evidence, clear procedure and fast appeal paths. Otherwise, safety regulation becomes a black box, and that weakens trust in rules that are actually needed.
Briefingshow
This is bigger than one model dispute. Washington is testing whether access to powerful AI systems themselves can be controlled, not just chips, data centers or training inputs. For companies, it means frontier models can suddenly become geopolitical goods, with export rules, customer cutoffs and compliance risk built into the product layer.
Sources