Anthropic Mythos exposed software flaws as Bank of America hired 1,000 contractors to patch vulnerabilities
Extra help for patching
Bank of America recruited 1,000 additional contractors after concerns arose about how quickly it could address vulnerabilities identified by Anthropic’s Mythos model. The CEO said Wednesday that the contractors would help manage the initial volume of work.
The contractors were brought in during the bank’s second quarter, adding to 3,000 existing technical staff. They remain at the bank, but their support is not expected to be needed over the longer term. The costs are also expected to decline as the immediate patching burden eases.
Wider concerns over Mythos
Concerns about handling the model’s discoveries were shared globally, following Anthropic’s presentation of Mythos as a powerful cybersecurity threat. Mozilla CTO Bobby Holley said in April that the model had found 271 flaws in Firefox 150.
Spending and AI forecasts
The CEO expects AI projects to deliver between $250 million and $300 million in benefits during 2026, rising to $700 million by the end of 2027. He did not explain how those estimates would be reached.
Bank of America CEO Moynihan did not detail how much of the “couple of hundred million plus” in remediation costs related to AI or Mythos, compared with work tied to the bank’s AML issues. He grouped the costs together.
Key points
- Bank of America added 1,000 contractors to help patch Mythos-discovered vulnerabilities.
- The contractors joined 3,000 technical staff and are expected to be less necessary over time.
- The CEO projected $250 million to $300 million in AI benefits for 2026.
- The bank did not separate Mythos remediation costs from other remediation spending.