JPMorgan CEO warns Anthropic's Mythos AI poses cybersecurity risks if widely available
JPMorgan CEO Jamie Dimon warned that Anthropic's Mythos AI model, which is highly effective at detecting software vulnerabilities, poses serious cybersecurity risks if made available to the general public. Dimon stated the risk is comparable to giving ballistic missiles to individuals, and the Bank of Spain has cautioned that the model could enable coordinated cyberattacks on the financial sector.
What happened
JPMorgan Chase CEO Jamie Dimon warned on Wednesday during Pennsylvania's Defense and Innovation Summit that Anthropic's Mythos AI model poses serious cybersecurity risks if made widely available to the public.
Dimon stated: 'Basically, with Mythos you're giving ballistic missiles to individuals.' Anthropic has reported that Mythos is so effective at detecting software vulnerabilities that it cannot be made available to the general public.
Reported impact
The Bank of Spain has warned that Mythos could generate 'more synchronized and extensive waves of cyberattacks' in the financial sector if it falls into the wrong hands.
Index analysis
JPMorgan has had access to Mythos since April and has been using it to test its own defenses and share information with suppliers and other companies. Following this, Dimon stated that his bank had hundreds of people working full-time to strengthen its systems.
Governance controls
Governments, large companies and strategic organizations have sought ways to access Mythos's capabilities under strict controls. Glaswing has become one of the main channels for testing and deploying this technology in high-security environments to strengthen cyber defense and anticipate increasingly sophisticated threats.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.