Particle.news
Download on the App Store

NYSE Says It Used Anthropic’s Mythos AI to Find and Fix Cyber Flaws

The disclosure prompts scrutiny over transparency and oversight of private AI programs that secure critical financial infrastructure.

Overview

  • NYSE President Lynn Martin told the House Financial Services Committee on Wednesday, Sept. 2, 2026, that the exchange participated in Anthropic’s Project Glasswing and used the company’s Mythos model to locate and remediate multiple cybersecurity vulnerabilities.
  • Project Glasswing is a restricted program that gives a small, vetted set of partners early access to Anthropic’s highest‑capability Mythos model for defensive testing and deep code analysis.
  • Martin declined to describe the specific flaws found or say when the fixes were applied, leaving technical details and remediation timing publicly unspecified.
  • Her testimony described proactive vulnerability testing rather than an active breach response, and neither Anthropic nor the NYSE offered additional public comment after the disclosure.
  • The revelation briefly coincided with a drop in Intercontinental Exchange shares and could prompt lawmakers and regulators to press for clearer disclosure rules and oversight of private AI tools used to protect market infrastructure.