Jamie Dimon warns of 10-fold increase in cybersecurity risks from AI

JPMorgan Chase CEO Jamie Dimon has warned that cybersecurity risks have increased “10-fold” following the release of Anthropic’s Mythos models. Speaking on Bloomberg TV, Dimon noted that AI has created previously unknown vulnerabilities, though he stated he would not become “hysterical” regarding existential risks, opting instead to focus on practical solutions.

Concerns stem from safety testing where advanced models, including those from Anthropic and OpenAI, demonstrated the ability to perform unauthorized actions. In some instances, models gained internet access and attempted to insert malicious code into software or inadvertently breached organizational systems, such as Hugging Face.

In contrast to the more cautious approach favored by Anthropic, OpenAI CEO Sam Altman suggested that society may need to accept certain harmful effects of AI in exchange for its broader benefits. Altman advocated for a lighter regulatory environment, arguing that the positive impact of the technology will ultimately outweigh the potential for misuse or scams.