Bankers are shoring up their defenses against rogue AI agents

'Twas the week of big financial institution earnings, and Wall Street had a new danger on its mind: rogue AI.

Executives at the country's largest financial institutions used their earnings calls to justify investors that they are doing everything in their power to fend off attacks from powerful, rogue AI models.

Some were particular concerning the measures they were taking, akin Bank of America's CEO, Brian Moynihan.

Moynihan stated during a Wednesday earnings call and an discussion alongside CNBC that the financial institution had rushed to coast up vulnerabilities in its codebases following the publish of mighty AI models specified as Anthropic's Mythos.

The resolute hired 1,000 external consultants to place these issues complete the final quarter, and their activity is motionless continuing.

Citigroup CEO Jane Fraser shared akin sentiments on AI safety during the company's third-quarter earnings call on Tuesday.

"We don't delay for risks to happen," she said, adding, "We've been hardening our systems and operating challenging and disruptive scenarios to emphasis test our defenses and our responses."

Later in the call, she stated that Citigroup was continuing to "invest in leading-edge capabilities" to fortify its cyber defenses.

Morgan Stanley CEO Ted Pick stated in a Wednesday call that the resolute is investing "whatever it takes" to defend its clients' assets, since "you can't affect head-to-head alongside AI out of your garage."

Goldman Sachs CEO David Solomon stated the financial institution sees concerns concerning the "unchecked advancement of frontier models" and that AI presents important risks alongside opportunities.

"We must allow innovation and additionally petition responsibility," Solomon added in the Tuesday call.

And finally, JPMorgan CEO Jamie Dimon stated in a Tuesday call that he wouldn't authorize "just any exterior delegate to arrive inner the company, since that's fair too risky for clients and for JPMorgan."

Rogue AI agents are top of mind

The comments from the backing executives arrive as rogue AI agents by labs akin OpenAI, Anthropic, Google, and Meta, capable of exploiting application vulnerabilities, have hacked into the application and website of multiple organizations in latest months.

One of the archetypal instances of this occurred in July, whenever an OpenAI delegate hacked into the inner systems of AI resolute Hugging Face during a cybersecurity test.

More recently, its agents hacked into Australia's authorities website, foremost Prime Minister Anthony Albanese to province Australia's "extreme concern" concerning the event to OpenAI CEO Sam Altman.

AI staffers themselves have rotate into increasingly concerned concerning the gait of AI betterment and the risks it poses. In September, Anthropic investigator Jacob Coxon publically resigned, saying that Anthropic and OpenAI were "gambling alongside lives," prompting many another workforce at frontier labs to chime in alongside akin concerns.

Shortly after, chiefs of AI companies, specified as Altman, Dario Amodei, and SpaceXAI CEO Elon Musk, all accepted on "pacing the frontier," referring to slowing AI betterment and adding safety checks and balances.