Bank of America had to hire 1,000 contractors to patch software flaws spotted by Anthropic Mythos
Bank also had to spent more than a hundred million dollars, but it's confident the spending will decrease in time.
On April 2026, Anthropic unveiled Mythos, a generative artificial intelligence (GenAI) model with advanced cybersecurity capabilities. The tool was so powerful, it could autonomously discover and exploit previously unknown software vulnerabilities. However, Anthropic later locked down access to Mythos, granting it to select organizations, including banks and cybersecurity firms.
Bank of America was among the organizations that received access to Mythos. The bank's CEO, Brian Moynihan, revealed that the AI model found numerous bugs in the system during the second quarter. In response, Bank of America hired 1,000 contractors and increased its workforce in the cybersecurity department to 3,000 people to address all the vulnerabilities.
Moynihan noted that the spending on fixing bugs and regulatory issues was "a couple hundred million plus" and expected to be temporary once the vulnerabilities were resolved.
The incident highlights the rapid progress of large language models and the heightened capabilities of AI in finding and exploiting software vulnerabilities. Other financial institutions, such as JPMorgan Chase and Goldman Sachs, have also ramped up their AI security efforts to address the threats exposed by Mythos. Nonetheless, the costs and resources required to keep up with these emerging vulnerabilities present a significant challenge for financial institutions.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
