Banks are among the most enthusiastic about AI, but they are also the most vulnerable to the technology’s growing cybersecurity threats. This vulnerability was brought into greater focus earlier this month with the release of Anthropic’s Mythos Preview, a highly advanced AI model that sparked concern across Wall Street. On earnings calls, JPMorgan Chase CEO Jamie Dimon and Goldman Sachs’ David Solomon said they were testing Mythos to better understand the new risks that come with rapid advances in artificial intelligence.
“AI has made it worse, made it more difficult,” Dimon told analysts today (April 14). “While we are trying to leverage AI, we are also very aware of the risks.”
These risks are key to Mythos, which Anthropic describes as too serious to be made public due to its ability to exploit vulnerabilities in critical software. Instead, the company invited a consortium of major firms, including JPMorgan, to test the model internally for use in strengthening their cybersecurity defences.
The inspection effort, dubbed Project Glasswing, takes its name from glasswing butterflies, which use transparent wings to hide in plain sight — a metaphor that Anthropic says reflects how hidden cyber vulnerabilities can evade detection. The initiative, which includes other Wall Street banks as well as Apple, Google and Nvidia, will be funded by $100 million in model usage credits from Anthropic.
Dimon said the launch of advanced models like Mythos has created “additional vulnerabilities” beyond banks. “Banks, of course, are tied to exchanges and all these other things that create other layers of risk. It’s a complex issue.”
Dimon, who led JPMorgan for two decades, stressed that cybersecurity remains a top priority for the country’s largest bank by assets and market capitalization. “We spend a lot of money. We have top experts. We are in constant contact with the government.”
Following the launch of the Mythos model, US Treasury Secretary Scott Besent and Federal Reserve Chairman Jerome Powell met with senior Wall Street executives in Washington, D.C., last week to discuss new threats posed by the model. While Damon was He was reportedly unable to attendJoining the meeting were peers including Brian Moynihan of Bank of America, Jane Fraser of Citigroup, Ted Beck of Morgan Stanley, and Charlie Scharf of Wells Fargo. Foreign central bank officials, including the Bank of Canada and the Bank of England, are hosting similar briefings with senior financial leaders.
Suleiman also attended the Treasury meeting. Goldman Sachs said during its bank’s earnings call yesterday (April 13). “It is clear that LLMs are making rapid progress,” Solomon told analysts. “We are fully aware of the improved capabilities of these new models.”
Despite their caution, both Dimon and Solomon remain confident that AI will eventually change banking for the better. JPMorgan has already applied AI to more than 500 use cases, while Goldman used it in programming and translation and earlier this year Partnership with Anthropy To integrate its CLOUD model across accounting and compliance.
“It’s not going to be a straight line when you have an acceleration of new technology,” Solomon said. “There will be bumps, there will be risk issues, there will be recalibrations.”
