Policy · 7h ago
Bank of America said Anthropic’s early-release Claude Mythos Preview found thousands of high-severity vulnerabilities in its systems, and the bank said it hired about 1,000 contractors and spent hundreds of millions to fix them. The disclosure landed as Bank of America also faced a $1.2 billion AML fine this week.
Anthropic told the U.S. House Financial Services Committee that Claude Mythos Preview had already surfaced thousands of high-severity issues, including some in every major operating system and web browser. In plain terms, the scan did not just flag a few bugs; it produced enough serious findings to turn remediation into a staffing and spending problem, not a routine patching task.
For large enterprises, especially banks, the point is that pre-release AI scanning can stress the remediation machine itself. If a tool finds flaws at this scale, the follow-on exposure is not only the software debt it reveals but the internal capacity needed to clear it.
2 sources covering this story
BofA spent millions on anti-money-laundering tech, still got fined $1.2B
It also hired 1,000 contractors to fix security bugs identified by Anthropic’s Claude Mythos.
Bank of America had to bring in 1,000 contractors to patch all the bugs Mythos identified
CEO said the workers were hired in Q2 and are still helping today
Part of the PlainSec briefing for 2026-10-16