Bank of America discloses on October 15 it added 1,000 contractors to patch vulnerabilities surfaced by Anthropic Mythos

Bank of America CEO Brian Moynihan told CNBC's Leslie Picker on October 15 US Eastern time that after frontier AI models including Anthropic Mythos surfaced security vulnerabilities in the bank's internal and vendor code, the bank hired 1,000 external contractors in Q2 to work alongside its existing 3,000-person patching team. Moynihan stressed that the risk was not limited to a single application or vendor and spanned a wide spectrum of internal and third-party code. On the analyst call Moynihan also disclosed that Q3 spending on software remediation and anti-money-laundering compliance combined for several hundred million dollars, without a separate breakdown for the Mythos-related portion. Bank of America shares fell as much as 8% intraday the same day. Moynihan said remediation spending should decline as patches are applied; the bank has about 130 internal AI projects, with AI-driven cost savings projected to grow from $300 million this year to about $700 million in 2027. Moynihan added that Anthropic's Mythos model had been made available in April to about 50 banks, tech companies and government agencies and that partner evaluations uncovered more than 10,000 high-severity vulnerabilities in important software, including a 27-year-old flaw in OpenBSD's firewall and a 16-year-old issue in FFmpeg.