Security Push Spurs AI Testing

The U.S. government is pushing financial firms to adopt advanced AI for cybersecurity, leading major banks to test Anthropic's Mythos model. The goal is to strengthen defenses against complex cyber threats by using AI to pinpoint system weaknesses. This push creates a complex dynamic between national security, business use, and the AI's powerful abilities.

Officials Urge Banks to Test AI for Vulnerabilities

Around April 7, 2026, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell encouraged leaders from major banks like JPMorgan Chase, Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley to use Mythos for finding vulnerabilities. This shows increasing government worry about broad cyber risks. Officials are asking these "systemically important" institutions to test the AI thoroughly on their systems. Mythos has already demonstrated strong offensive and defensive cyber skills by finding and exploiting chained vulnerabilities in web browsers, something human analysts often miss. This means the AI is a powerful, but potentially risky, tool for improving financial sector security.

AI's Dual Role in Cybersecurity Finance

Mythos is a major step forward for AI in cybersecurity. AI spending in financial services is expected to hit $75.19 billion in 2026. Around April 7, 2026, Anthropic launched 'Project Glasswing', uniting tech giants like Amazon Web Services, Apple, Google, and Microsoft to secure software infrastructure with Mythos. This group aims to tackle threats before they can be used against institutions. The U.S. Treasury Department is also creating a guide for AI in finance, releasing an AI Lexicon and a Financial Services AI Risk Management Framework (FS AI RMF) in February 2026. These aim to create common terms and risk management rules for responsible AI use, improving security and stability.

AI's Exploitation Risk and Anthropic's Legal Battles

Anthropic's legal problems temper government interest in Mythos. The Pentagon labeled Anthropic a 'supply-chain risk' over ideological disagreements, sparking a legal fight. On March 26, 2026, a California federal court blocked a Pentagon order. However, a federal appeals court in Washington D.C. recently denied Anthropic's request to halt the designation during the legal process, creating judicial uncertainty. The AI's offensive skills, useful for defense, also carry a risk of exploitation. Its ability to chain multiple vulnerabilities—a complex hacking tactic—raises concerns. The chance of unintended consequences or misuse, combined with Anthropic's difficult ties with some government bodies, poses a significant risk for banks using its tech. As of April 10, 2026, major banks like JPMorgan Chase (market cap ~$837.7B, forward P/E 13.47x) and Goldman Sachs (market cap ~$267.37B, P/E 17.67x) trade near yearly highs, showing investor confidence. However, the challenge of balancing regulation and ethics with advanced AI like Mythos creates an unpriced risk.

Financial Sector's AI Future

The financial sector is expected to use AI for cybersecurity more, driven by regulatory backing and changing threats. Successfully using tools like Mythos will depend on strong governance and ongoing risk checks. Upcoming earnings reports from banks like Citigroup (expected 34% EPS growth) and Goldman Sachs (expected $16.35 EPS) will offer more details on the sector's financial strength and its ability to invest in and manage advanced AI during economic shifts.