Abstract:
Bank of America has hired a slew of outside contractors to help patch software vulnerabilities exposed by advanced artificial intelligence models, including Anthropic’s Mythos. In an interview after the bank released its third-quarter earnings report, CEO Moynihan said that earlier this year, the company mobilized thousands of employees and external personnel to urgently harden its extensive internal and vendor code.
"That risk is real, which is why when the leading-edge models emerged, a lot of us sat back and thought, how do we get through this quickly?" Moynihan said.
"We added 1,000 contractors in the second quarter to help us with patch repairs; those folks are still here today, on top of the 3,000 we already have in the field, continuing to help us," he said.
Moynihan told analysts earlier on Wednesday that the bank spent "more than $200 million" in the third quarter on software fixes and regulatory issues related to anti-money laundering compliance.
These remarks are the first time a major U.S. company has specified the amount of manpower and money it has invested in protecting its software systems from Mythos. Mythos was released to a small group of banks, technology companies and government agencies in April.
They also highlight the double-edged sword nature of AI: While AI can help companies save money by increasing productivity, models like Mythos' ability to find and exploit decades-old software vulnerabilities also forces companies to undertake costly defensive responses.
“We have to stay ahead of the curve, and we have to realize that the people doing software development are the ones trying to sell us services,” Moynihan said.
However, Moynihan - speaking as the company's stock price fell as much as 8% on Wednesday - sought to reassure investors that the increase in spending was temporary.
"Once you've patched it, you don't have to patch it as many times... the process will be phased out," Moynihan said, adding that the bank expects the cost of remediation to fall over time.
In other ways, AI is helping reduce costs. Moynihan told analysts that Bank of America has about 130 internal AI projects and the savings from these projects are expected to increase from $300 million this year to $700 million in 2027.
He said the total number of Bank of America employees has decreased by about 3,400 over the past year as the company laid off about 5% of its managers. It's unclear how much of this headcount reduction can be attributed to AI.
Comments