Overview
- JPMorgan CEO Jamie Dimon said advanced frontier models like Anthropic’s Mythos have exposed previously unknown software vulnerabilities and that cyber risk has risen sharply since Mythos’s emergence.
- Dimon acknowledged these models can help defenders find and patch old, hidden flaws but warned the same capability can empower attackers if misused.
- He framed the response as pragmatic work rather than panic and urged banks and developers to implement practical safeguards to reduce exploitability.
- Recent safety-testing disclosures from Anthropic and OpenAI have intensified pressure on industry leaders and some lawmakers to pursue stronger oversight and cross-border coordination of frontier AI.
- Banks continue to expand AI use for business purposes even as they adjust operations and staffing, creating trade-offs between efficiency gains and elevated systemwide security exposure.