Bank of America Joins Project Glasswing to Test Anthropic’s Mythos AI for Cybersecurity
Bank of America is among four systemically important banks testing Anthropic PBC’s Mythos AI model internally to detect cybersecurity vulnerabilities following a US Treasury and Federal Reserve summit. Mythos access is part of “Project Glasswing,” limited to select institutions to bolster defenses against advanced threats.
1. Mythos Model Testing at Bank of America
Bank of America has begun internal testing of Anthropic PBC’s Mythos AI model to identify and mitigate cybersecurity vulnerabilities. The model’s offensive and defensive capabilities are evaluated against simulated threats to strengthen system-wide defenses.
2. Summit with US Treasury and Fed
A recent summit convened by Treasury and Federal Reserve officials urged systemically important banks to adopt Mythos internally. Attendees were advised to leverage the model proactively to uncover hidden network vulnerabilities before potential cyberattacks.
3. Project Glasswing’s Limited Release
“Project Glasswing” restricts Mythos access to a select group of critical institutions, including Bank of America, JPMorgan Chase, Goldman Sachs, Citigroup and Morgan Stanley. This phased rollout aims to secure essential banking infrastructures against emerging AI-driven threats.