Bank of America brought in 1,000 contractors to patch AI-exposed flaws
Bank of America chief executive Brian Moynihan says about 1,000 outside contractors joined an existing internal team to repair software weaknesses exposed by AI models, including Anthropic’s Mythos. The consultants sit outside the bank’s employee headcount. His disclosure gives a staffing picture of the repair work that follows automated vulnerability discovery.
Artificial Intelligence··Evening
Contractors joined an existing repair team
Bank of America chief executive Brian Moynihan said about 1,000 contractors were hired in the second quarter to patch software weaknesses exposed by AI models, including Anthropic’s Mythos. Around 3,000 bank employees were already working on the task.[1], [2]
Moynihan disclosed the staffing in an October 14 CNBC interview. The work covered software used by the bank and code supplied by outside vendors. The Next Web attributes the figures to the chief executive and says it has not independently verified them.[1]
Consultants sit outside the employee count
Moynihan explicitly distinguished the extra consultants from employees: they were not included in corporate headcount. The bank’s employee total therefore leaves out this additional repair workforce. The disclosure provides no contractors’ pay, contract durations or hours worked by either group.[1]
Restricted access to Mythos preceded the repairs
Anthropic made Mythos available in April through Project Glasswing, a restricted programme for companies, banks and government agencies to find and patch vulnerabilities. Moynihan had said in July that such tools increased the work banks faced. His quarterly remediation spending estimate also included anti-money-laundering controls, so the whole amount cannot be assigned to AI-related repairs.[1]
Related columns
For more information on this topic, you can read the related columns.
