JPMorgan’s Jamie Dimon Says AI Cyber Risks Rose Tenfold After Mythos
JPMorgan CEO Jamie Dimon says cybersecurity risks increased roughly tenfold after Anthropic’s Mythos AI model emerged, warning that increasingly capable AI systems are creating vulnerabilities that were previously difficult to identify.
JPMorgan Chase CEO Jamie Dimon has issued a fresh warning about the cybersecurity implications of increasingly powerful artificial intelligence systems, saying the risks associated with AI have risen dramatically following the emergence of Anthropic’s Mythos model. Speaking to Bloomberg TV on October 6, Dimon said cyber risks had increased by roughly tenfold after Mythos, adding that AI had created vulnerabilities that companies had not previously known existed.
Dimon’s concerns are linked to Mythos’ ability to identify software vulnerabilities at an advanced level. Anthropic has not broadly released the system, instead providing access to selected organizations and security researchers because of the potential risks associated with its capabilities. JPMorgan is among the organizations working with Anthropic on testing the technology, giving the bank an opportunity to use advanced AI to identify weaknesses while also evaluating the security challenges that such systems could create.
The JPMorgan chief has previously expressed similar concerns about unrestricted access to Mythos. In July, Dimon described the model as a serious issue and compared widespread access to its capabilities to giving individuals extremely powerful weapons. At the time, he said governments needed to control access to advanced AI because systems capable of finding software vulnerabilities could potentially be misused by criminals, hostile organizations or foreign intelligence services.
Despite the warning, Dimon said he does not believe the appropriate response is to become hysterical about whether AI represents an existential threat. Instead, he argued that businesses and governments should focus on understanding the technology and building stronger defenses. He also acknowledged that AI could produce enormous benefits, including advances in medicine, materials science and other areas, meaning the technology’s potential advantages need to be considered alongside its security risks.
Dimon’s comments underline a growing dilemma for banks and other critical institutions: the same AI systems that can help organizations discover vulnerabilities faster could potentially give attackers similar capabilities. As AI agents become more autonomous and capable of operating across computer systems, cybersecurity teams are likely to face a constantly changing threat landscape. For JPMorgan, which has already invested heavily in technology and AI, the challenge will be to capture the productivity and security benefits of advanced models while ensuring that their capabilities do not create vulnerabilities greater than the problems they are designed to solve.
