Anthropic CEO warns of a cyber danger window as AI finds vulnerabilities
AI cyber defense helps fix software flaws before attackers exploit them. See how Anthropic’s Mythos findings raise the stakes for banks and businesses.
Anthropic CEO Dario Amodei said the rapid advance of AI has created a limited window for companies, governments, and banks to fix software weaknesses uncovered by the company’s Mythos model.
Speaking at an Anthropic financial services event in New York on Tuesday, Amodei said the model has identified tens of thousands of vulnerabilities across software systems, including longstanding bugs in critical products. He warned that adversaries could exploit these issues if they remain unpatched.
Amodei said competing AI systems in China may be only six to 12 months behind Anthropic’s capabilities, which he described as adding urgency to the need for stronger cyber defenses. JPMorgan Chase CEO Jamie Dimon joined him at the event, where Anthropic also announced new AI agents for banking and backoffice tasks and a broader Microsoft Office integration.