Bank of America has engaged approximately 1,000 external contractors to assist in repairing software security vulnerabilities exposed by cutting-edge artificial intelligence models, including Anthropic's AI model Mythos. Following the release of third-quarter financial results, the bank's CEO, Moynihan, stated that thousands of internal employees and external professionals were mobilized earlier this year to urgently strengthen the security of internal and supplier code. This repair effort involves collaboration between the existing 3,000 internal employees and the newly added 1,000 contractors to address deep-seated vulnerabilities identified by AI models. Moynihan noted that while repair costs are higher in the short term, related expenses are expected to decline as vulnerabilities are progressively resolved. Additionally, Bank of America has approximately 130 AI projects internally, with estimated cost savings projected to increase from $300 million this year to $700 million by 2027.