Anthropic blocks UK AI security institute testing Claude Mythos 5.1
Anthropic, a major AI company, just told the U.K.'s top AI Security Institute (AISI) they can't test its newest model, Claude Mythos 5.1, marking the first time AISI has been shut out like this.
The move comes as US export controls on advanced tech get tighter, sparking worries that international teamwork and oversight on powerful AI could take a hit.
UK officials cite Trump-era export rules
Before launch, Anthropic only shared Claude Mythos 5.1 with select U.S.-based organizations.
UK officials aren't thrilled and see this as a sign that American rules might start blocking other countries from cutting-edge AI tools.
These restrictions date back to Trump-era policies limiting foreign access to advanced US systems, and allies like the UK are now wondering how much global cooperation will be possible going forward.