Anthropic’s Mythos model raises cybersecurity concerns for banking sector

Experts warned that Anthropic's advanced Mythos model, capable of detecting and exploiting software vulnerabilities, could enable large-scale AI-driven cyberattacks on legacy banking systems. Anthropic has limited the model to controlled preview via Project Glasswing. Reported in Economic Times EnterpriseAI on April 14. It matters because it highlights immediate regulatory and risk-management implications for critical financial infrastructure when frontier AI coding capabilities are deployed.