Bank of America Hires 1,000 Contractors to Patch AI-Exposed Software Flaws

Bank of America has retained 1,000 external contractors to remediate software vulnerabilities identified by advanced artificial intelligence models, including Anthropic's Mythos, according to statements made by the bank's chief executive, Brian Moynihan. Speaking to CNBC on Wednesday following the release of third-quarter earnings, Moynihan detailed the immediate mobilization of resources required to address security risks uncovered by frontier AI capabilities.

Moynihan explained that the bank activated a large-scale response earlier this year to secure both internal code and vendor-supplied systems. He noted that the 1,000 contractors were added during the second quarter to support a team of 3,000 existing personnel dedicated to these technical areas. The additional workforce remains in place as part of the ongoing defensive effort.

The financial impact of these remediation efforts is substantial. Moynihan stated that the bank spent more than a couple hundred million dollars in the third quarter on software fixes and regulatory issues related to anti-money laundering compliance. These remarks represent some of the first specific disclosures by a major U.S. corporation regarding the personnel and capital required to defend against vulnerabilities exposed by Mythos, a model released to a limited group of banks, technology firms, and government agencies in April.

The situation illustrates the dual nature of AI adoption. While generative models can enhance productivity and reduce costs, their ability to identify and exploit long-standing software defects has necessitated expensive defensive measures. Moynihan highlighted the complexity of this dynamic, noting that the entities developing the software are often the same ones selling services to the bank. "We've got to stay ahead of them," he said.

Despite the significant expenditure, Moynihan sought to reassure investors that the increased costs are temporary. He indicated that remediation costs are expected to decline over time as initial patches are applied, describing the process as a pipeline that moves through once vulnerabilities are addressed. This reassurance came as Bank of America shares fell by as much as 8% on Wednesday.

Conversely, AI is contributing to cost reductions in other areas of the business. The bank currently operates approximately 130 internal AI projects. Moynihan projected that savings from these initiatives will increase from $300 million this year to $700 million by 2027. Additionally, the bank's headcount decreased by roughly 3,400 employees over the past year, including a cut of about 5% of managers. Moynihan did not specify how much of this workforce reduction is directly attributable to AI integration.