Anthropic's Powerful AI Model Sparks Emergency Banking Meetings

US Treasury and Federal Reserve warn of systemic cybersecurity risks from Anthropic's Claude Mythos AI

Apr. 12, 2026 at 2:35pm

Got story updates? Submit your updates here. ›

A highly detailed, glowing 3D illustration of a complex network of interconnected digital circuits and data streams, representing the underlying infrastructure of the global financial system. Pulsing neon cyan and magenta lights convey a sense of both power and vulnerability in the face of emerging AI threats.Advanced AI models like Anthropic's Claude Mythos Preview expose critical vulnerabilities in the digital foundations of the global financial system.Washington Today

U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an urgent meeting with the chief executives of the largest American banks to warn of systemic cybersecurity risks posed by Anthropic's latest artificial intelligence model, Claude Mythos Preview. The emergency huddle in Washington included leaders from Goldman Sachs, Wells Fargo, Morgan Stanley, Bank of America, and Citigroup, focusing on the potential for AI-driven cyberattacks to undermine the global financial system.

Why it matters

The warnings center on Claude Mythos Preview's ability to identify software vulnerabilities and exploit zero-day flaws, raising concerns that the technology could be used to target banking platforms, potentially forcing them offline or allowing unauthorized actors to tamper with financial data. Government officials expressed alarm over the most severe scenario - the possibility of account balances being digitally wiped or reduced to zero.

The details

To mitigate the risk of widespread exposure, Anthropic is limiting the distribution of Claude Mythos Preview, intending to offer the model to only a few dozen companies. The urgency of the meeting reflects a shift in concern from individual job displacement to systemic infrastructure failure, as advanced AI systems could potentially bypass existing corporate defenses to destabilize financial institutions.

  • The emergency meeting took place during the week of April 6, 2026.

The players

Scott Bessent

U.S. Treasury Secretary.

Jerome Powell

Federal Reserve Chair.

Anthropic

An artificial intelligence company that developed the Claude Mythos Preview model.

Goldman Sachs

One of the largest American banks that participated in the emergency meeting.

Wells Fargo

One of the largest American banks that participated in the emergency meeting.

Got photos? Submit your photos here. ›

What’s next

Anthropic is limiting the distribution of Claude Mythos Preview to only a few dozen companies in order to mitigate the risk of widespread exposure.

The takeaway

The emergency meeting between government officials and banking leaders highlights the growing concerns around the potential for advanced AI systems to undermine the stability of the global financial system. As AI capabilities continue to evolve, there are increasing worries about the systemic risks posed by technologies that can exploit vulnerabilities and disrupt critical infrastructure.