Bank of America Adds 1,000 Contractors to Fix AI‑Exposed Software Flaws

Bank of America announced that it has hired about 1,000 outside contractors to help patch software weaknesses that were revealed by Anthropic’s Mythos AI model. The move follows the bank’s own 3,000 employees who were already working on the same remediation effort. According to CEO Brian Moynihan, the cost of fixing these flaws and repairing anti‑money‑laundering controls was “a couple hundred million dollars plus” for the quarter. Moynihan said the contractors are not counted in the bank’s headcount because they are hired as consultants.

The remediation effort involves both proprietary and vendor software, and the bank has been working with the contractors since the second quarter. The internal staff and the contractors are expected to reduce the Mythos‑related workload as the patching proceeds. The total cost for the quarter was reported to be in the hundreds of millions, a figure that will likely drop as the additional workers take on more of the load. This expense comes on top of the bank’s regular operations and has already impacted earnings.

Regulatory pressure has also been a factor in the bank’s decision. Treasury announced penalties of more than $1.21 billion on 13 October for failing to report suspicious transactions worth $5 billion. U.S. regulators paused some cyber examinations to give banks time to address the vulnerabilities identified by Mythos. The European Central Bank has requested plans to mitigate AI‑enabled threats by the end of October, further increasing scrutiny.

Looking ahead, Moynihan expects the costs associated with the Mythos patching to decline over time. The bank’s 130 internal AI projects are projected to generate about $700 million in savings next year, offsetting some of the short‑term expenses. The company also indicated that the Mythos work will shrink now that the additional workers are in place.