Bank of America has hired 1,000 outside contractors to fix software security flaws found by advanced artificial intelligence (AI) models, including Anthropic’s Mythos. The move shows how AI tools can identify vulnerabilities faster than companies can fix them.
Bank of America CEO Brian Moynihan revealed the additional hiring in October 2026. The new contractors joined around 3,000 existing employees at the bank. Their work involves finding and fixing security flaws in Bank of America’s systems and third-party software.
The bank began hiring the contractors in the second quarter of 2026. They were still working with the company in October as Bank of America continued to address vulnerabilities identified by AI models.
Why Bank of America Hired 1,000 Contractors
Bank of America hired additional contractors after advanced AI models found software vulnerabilities that needed further checks and repairs. Anthropic’s Mythos can identify software flaws and assess how attackers could exploit them.
These AI capabilities can help security teams find weaknesses that traditional testing methods may miss. However, finding a vulnerability is only the first step. Teams must examine each issue, decide which ones need urgent attention and fix them without disrupting banking services.
Bank of America needed more technical staff to handle the growing workload. The 1,000 contractors are helping existing employees fix vulnerabilities in the bank’s software and applications developed by third-party vendors.
ALSO READ: AI Cyber Attacks Statistics [2025 and beyond]
Anthropic’s Mythos Raises Cybersecurity Concerns
Anthropic introduced Mythos in April 2026 with advanced capabilities to identify software vulnerabilities. The company limited access to the model because its ability to find and exploit security flaws could also create risks if attackers used it for cyberattacks.
Mythos can analyse software code, detect weaknesses and identify ways to exploit them. This creates a challenge for businesses that must fix security flaws before cybercriminals can take advantage of them.
To address these risks, Anthropic launched Project Glasswing. The initiative gives selected organisations access to its cybersecurity tools to help them find and fix weaknesses in critical software.
Banks face particular pressure to address these security risks. Financial institutions rely on complex IT systems and handle sensitive customer data. Any unpatched vulnerability could expose these systems to potential attacks.
Reuters reported in May 2026 that US banks were working to fix IT vulnerabilities identified by Mythos. This has raised questions about whether financial institutions could repair these weaknesses quickly enough. Advanced AI tools have made it easier to discover software flaws.
How AI-Detected Security Flaws Increase Cybersecurity Costs
Identifying software vulnerabilities is only the first step in cybersecurity. Once a flaw is identified, security teams must assess its severity, understand the risks and find a way to fix it.
The teams must then test the fixes to make sure they do not create new security problems or disrupt existing services. Large banks also need to coordinate repairs across their own systems and software supplied by external vendors.
This work can take significant time and resources, particularly when AI tools uncover large numbers of vulnerabilities across extensive amounts of code.
Bank of America’s decision to hire 1,000 contractors shows how large-scale AI security testing can increase the demand for technical staff. The additional workers are helping the bank investigate and fix vulnerabilities while its existing teams continue their work.
During discussions about the bank’s third-quarter results, CEO Brian Moynihan also addressed the costs involved in fixing these issues. However, the reported spending covered other areas, including anti-money-laundering matters. Therefore, it should not be treated as the confirmed cost of fixing vulnerabilities identified by AI models alone.
Bank of America Highlights the Challenges of AI Cybersecurity
Bank of America’s decision highlights a challenge facing businesses that use advanced AI cybersecurity tools. These tools can find software vulnerabilities quickly. But companies also need enough skilled workers and resources to fix the problems.
For banks, delays in fixing serious security flaws can increase the risk of cyberattacks, data leaks and service disruptions. Financial institutions may need to invest in automated security testing. They may also need to strengthen the teams responsible for fixing the vulnerabilities these tools uncover.
AI tools may eventually help security teams decide which vulnerabilities to fix first and automate some repairs. However, experts still need to test these fixes carefully. Mistakes in banking systems could affect customers and disrupt essential financial services.
ALSO READ: OpenAI Agent Accessed Non-Public Files on Australian Government Medicare Portal
Bank of America’s experience shows why companies must prepare for the workload created by advanced AI security tools. These tools can find software weaknesses faster. However, existing teams may struggle to fix all the problems they uncover.
As AI cybersecurity tools improve, businesses will need to strengthen their ability to identify and assess software vulnerabilities. They must also ensure they have enough staff to fix these issues.
