Bank of America mobilized a large remediation effort after advanced AI models including Anthropic’s Mythos revealed software vulnerabilities. CEO Brian Moynihan said the bank brought on 1,000 outside contractors to supplement roughly 3,000 in-house staff to patch proprietary and vendor code, and that third‑quarter remediation plus anti‑money‑laundering compliance added “a couple hundred million plus” to costs. Mythos, released in April to a limited set of banks, tech firms and agencies, exposed decades‑old flaws and forced a rapid, costly defensive response that coincided with an up-to-8% dip in the bank’s shares.
Moynihan framed the spending as temporary and necessary, arguing that once systems are patched the pipeline of remediation will ease. At the same time the bank is accelerating AI deployment: about 130 internal AI projects are projected to increase savings from $300 million this year to $700 million in 2027. The firm also reduced headcount by roughly 3,400 employees (about 5% of managers) over the past year, with unclear overlap between layoffs and automation. The disclosure highlights the double‑edged nature of frontier AI - prompting expensive defensive work even as it powers future productivity gains.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.