Anthropic Mythos reveals ‘more vulnerabilities’ for cyberattacks
Jamie Dimon (R), chief executive officer of JPMorgan Chase & Co., leaves the U.S. Capitol on Wednesday, February 25, 2026 in Washington, DC, USA.
Graeme Sloan | Bloomberg | Getty Images
JPMorgan Chase AI tools can help companies protect themselves against cyberattacks, but they primarily make them more vulnerable, CEO Jamie Dimon said Tuesday.
Dimon said JPMorgan is testing a preview of Anthropic’s latest model, the Mythos. announced by the AI firm last week — as part of a broader effort to harness the benefits of artificial intelligence while protecting against bad actors using the same technology.
“AI has made it worse, made it more difficult,” Dimon told analysts on the bank’s earnings call Tuesday morning. “It creates additional vulnerabilities and maybe creates better ways to empower yourself in the future.”
Asked by a reporter about Mythos, Dimon appeared to reference Anthropic’s warning that the model had already found thousands of vulnerabilities in enterprise software.
“I think you read exactly what happened,” Dimon said. “This shows that many more vulnerabilities need to be fixed.”
The revelations reveal how artificial intelligence, a technology welcomed by companies for productivity, has turned into a serious threat by offering bad actors new ways to infiltrate technology systems. Last week, Treasury Secretary Scott Bessent called a meeting of bank CEOs to discuss the risks posed by Mythos.
Dimon said JPMorgan, the world’s largest bank by market value, has been investing heavily for years with dedicated teams and constant coordination with government agencies to stay ahead of threats.
“We spend a lot of money. We have high-level experts. We are in constant contact with the government,” he said. “This is a full-time job and we do it all the time.”
‘Attack mode’
Still, the CEO warned that the risks extend beyond any single institution, given the interconnected nature of the financial system.
“That doesn’t mean that everything that banks rely on is that well protected,” Dimon said. “Banks are tied to stock markets and other things that create other layers of risk.”
JPMorgan Chief Financial Officer Jeremy Barnum said the industry has long been aware that AI cuts both ways in cybersecurity.
“These tools can make it easier to find vulnerabilities, but they can also potentially be used in attack mode by bad actors,” Barnum said on the earnings call. Recent advances by Anthropic and others intensify an existing trend, he said.
Dimon also said that while advanced artificial intelligence tools are important, old-school cybersecurity practices are still necessary.
“A big part of this is about hygiene… how do you protect your data? How do you protect your networks, your routers, your hardware, by changing your password?” he said. “Doing all these things correctly reduces risk significantly.”
Goldman Sachs CEO David Solomon said during an earnings call Monday that his bank was testing Mythos but declined to comment further.
