Anthropic has granted the European Union access to its Mythos cybersecurity AI model, marking the first time the powerful tool has been offered outside the US and UK. The EU's cyber security agency Enisa confirmed negotiations are underway, though conditions for access remain undefined. European Commission officials visited San Francisco last week to discuss joining Project Glasswing, an industry coalition using the model to identify security vulnerabilities.

Mythos initially reached only select critical infrastructure companies including Microsoft, Apple, JPMorgan, and CrowdStrike due to its advanced capabilities. UK government stands as the sole entity outside the US to have gained access through its AI Security Institute. Anthropic CEO Dario Amodei previously expressed willingness to share the technology with US-allied governments to defend democracies under attack, while cautioning against potential misuse.

The exact terms of EU access still require agreement, including how much visibility Anthropic would gain into EU systems. Anthropic declined to comment on the negotiations but announced plans to release Mythos to all customers in coming weeks after developing stronger cyber safeguards. The discussions represent a significant expansion of the AI cybersecurity model's international reach.