Citigroup Joins Five Banks Testing Anthropic’s Mythos AI Under Treasury Push
US Treasury Secretary Scott Bessent and Fed Chair Jerome Powell summoned Citigroup among five systemically important banks on April 7 to deploy Anthropic’s Mythos AI model for internal vulnerability scanning. Mythos, released to a few dozen firms under Project Glasswing, can detect and potentially exploit web browser vulnerabilities.
1. US Regulators Encourage AI Testing
US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened five systemically important banks, including Citigroup, on April 7 at Treasury headquarters to urge internal deployment of Anthropic’s Mythos model for proactive vulnerability detection.
2. Banks Participating in Mythos Trials
In addition to Citigroup, JPMorgan Chase, Goldman Sachs, Bank of America and Morgan Stanley have gained access to Mythos for internal cybersecurity testing, with banks either testing in-house or awaiting rollout under the initiative.
3. Anthropic’s Project Glasswing and Capabilities
Anthropic has limited Mythos release to a few dozen firms under Project Glasswing, highlighting the model’s ability to identify and potentially exploit vulnerabilities such as web browser cross-site data exposures to strengthen defenses.