Bank of America Brings in 1,000 Contractors to Patch Vulnerabilities Found by Anthropic’s Mythos as Remediation Costs Reach Hundreds of Millions

https://www.theregister.com/security/2026/10/15/bank-of-america-had-to-bring-in-1000-contractors-to-patch-all-the-bugs-mythos-identified/5303275

Bank of America recruited an additional 1,000 contractors during the second quarter of 2026 to help patch the volume of vulnerabilities discovered by Anthropic’s Mythos AI security model, with chairman and CEO confirming during the company’s third quarter earnings call that combined remediation costs related to Mythos discoveries and concurrent anti-money laundering orders had reached “a couple of hundred million plus.” The contractors were brought in on top of the bank’s existing 3,000 technical staff after concerns grew internally over how quickly the organisation could work through the volume of flaws Mythos had identified, a challenge shared broadly across early adopters of the model given how Anthropic had marketed it as a transformative and all-powerful capability that threatened the existing status quo of cybersecurity. The contractors currently remain in place but that their assistance would not be needed in the longer term as the initial remediation burden eases, and that the associated costs are expected to decline over time. The CEO told CNBC that each of the approximately 130 AI-related projects the bank currently has running is expected to achieve a two to three times return on investment over time, with AI projects projected to deliver between $250 million and $300 million in benefits in 2026 alone, rising to $700 million by the end of 2027, while also noting that AI has made the bank’s technology engineering workforce approximately ten percent more efficient.