JPMorgan CEO Says AI-Related Cyber Risks Rose Sharply

JPMorgan CEO Jamie Dimon said Oct. 6 that artificial intelligence had sharply increased cybersecurity risks, calling cyber the bank’s biggest risk as AI becomes more deeply tied to markets and corporate spending.

“Risks from AI went up 10-fold after Mythos,” Dimon said. “AI created vulnerabilities that we didn’t know about.”

The model he referenced is Claude Mythos Preview, which Anthropic introduced in April as a restricted model for defensive cybersecurity work. JPMorgan was among the launch partners in Anthropic’s Project Glasswing initiative.

Anthropic said roughly 50 initial partners used Mythos Preview to identify more than 10,000 high- or critical-severity vulnerabilities. The program planned to add approximately 150 organizations in more than 15 countries.

Anthropic also warned that attackers could gain an advantage if frontier artificial intelligence companies are careless when releasing models. The warning highlights the dual-use risk of systems designed to find security flaws: The same capabilities that help defenders can also expose new tools to attackers.

The cyber warning comes as AI remains closely linked to U.S. equity markets. Nearly 40% of the S&P 500’s market capitalization was directly affected by AI use, investment, infrastructure construction and productivity gains.

JPMorgan’s 2026 company update put the bank’s annual technology spending at approximately $20 billion. JPMorgan’s update did not break out AI spending separately.

The bank has also discussed the difficulty of measuring returns from AI investment. In June 2025, Dimon said JPMorgan was receiving about $2 billion in benefits from roughly $2 billion in AI spending, while CFO Jeremy Barnum said proving those gains was difficult. JPMorgan’s AI savings claims have been difficult to measure.