UK financial regulators assess risks from Anthropic’s latest AI model

British financial regulators held urgent talks with the National Cyber Security Centre and major banks to evaluate risks posed by Anthropic's latest unreleased AI model (Claude Mythos Preview), deployed only via Project Glasswing for defensive cybersecurity. Reported across Indian business media on April 12. It matters because it demonstrates proactive regulatory response to frontier AI capabilities that could expose vulnerabilities in critical financial IT systems.