Bank of America told investors it brought on about 1,000 external contractors in the second quarter of 2026 to patch software bugs that Anthropic's Mythos AI model exposed. The work adds to a permanent team of roughly 3,000 staff in the bank’s security and remediation group. The bank said the effort cost “a couple hundred million plus” in the third quarter.
CEO Brian Moynihan said the AI‑driven risk was real and forced the bank into a rapid defensive sprint. While the spending spiked this year, he expects remediation costs to fall once the most urgent patches are applied.
Why the AI model mattered
Mythos, released to a limited set of banks, tech firms, and government agencies in April 2026, can scan code and point out weaknesses that have existed for decades. Those findings gave attackers new ways to exploit legacy systems, prompting the bank to act fast.
Broader AI impact at the bank
The same CEO noted that Bank of America runs about 130 internal AI projects, which are projected to generate $700 million in savings by 2027, up from $300 million this year. The firm also cut roughly 3,400 jobs, including a 5% reduction in managers, though it did not link those cuts directly to AI.
Why it matters
For everyday customers, the patching effort means the bank’s systems should stay more secure against new AI‑powered attacks. However, the high short‑term cost shows that adopting advanced AI can bring unexpected expenses as companies scramble to fix hidden vulnerabilities.
