Dimon Warns Powerful AI Models Could Put 'Ballistic Missiles' In Individuals' Hands
By CU Today Staff —
NEW YORK—JPMorgan Chase CEO Jamie Dimon warned that broad public access to Anthropic's most advanced Mythos-class artificial intelligence models poses significant national security risks, saying the technology is powerful enough that releasing it without proper controls would be akin to giving "ballistic missiles to individuals," Business Insider reported.
Speaking at the Pennsylvania Defense and Innovation Summit, Dimon said Mythos is "a real issue" and praised the U.S. government's oversight of the technology, saying officials are working to ensure the models can be controlled before wider deployment. He made the remarks during an "Investing in America" panel alongside Sen. Dave McCormick, according to Business Insider.
Business Insider reported that Anthropic has acknowledged the cybersecurity risks posed by its flagship Mythos 5 model, which demonstrated an exceptional ability to identify high-severity software vulnerabilities. The company initially withheld the model from broad release, limiting access to select U.S. organizations, before later introducing Fable 5, a related model with additional safety guardrails.
Business Insider said the U.S. government temporarily imposed export controls on both Mythos 5 and Fable 5 after determining the models' safeguards could potentially be bypassed. Anthropic subsequently suspended access before the Commerce Department lifted the restrictions on June 30, allowing the company to begin restoring availability.
Originally reported by CU Today.