La présidente du NYSE affirme que la Bourse a utilisé l’IA d’Anthropic pour corriger des failles de cybersécurité
Points clés
- •Lynn Martin a déclaré que le NYSE avait utilisé Project Glasswing d’Anthropic pour identifier et corriger plusieurs faiblesses de cybersécurité.
- •L’information a été divulguée lors d’un témoignage devant la commission des services financiers de la Chambre.
- •Elle n’a pas révélé quelles étaient les vulnérabilités, quand elles ont été corrigées ni quelles ressources ont été nécessaires.
- •Project Glasswing donne à un nombre limité d’organisations vérifiées un accès anticipé au modèle Mythos d’Anthropic pour des tests de sécurité défensifs.
- •L’action Intercontinental Exchange a chuté de 1.39% le jour du témoignage.
NYSE President Lynn Martin told Congress that the New York Stock Exchange used Anthropic’s Project Glasswing to identify cybersecurity vulnerabilities and resolve them quickly.
Martin made the disclosure on Wednesday while testifying before the House Financial Services Committee and answering questions from Representative Bill Huizenga, a Republican from Michigan, about preparations for coordinated cyberattacks. She said the exchange participated in Project Glasswing as an early partner and that the technology helped it locate multiple issues in its systems.
Martin did not disclose what the vulnerabilities were or when they were fixed.
What Project Glasswing Is
Project Glasswing is Anthropic’s program that provides a small group of vetted organizations with early access to its Mythos AI model for defensive security testing. Anthropic describes Mythos as its most capable model tier, and access is currently limited to selected partners rather than the public.
According to Martin, the NYSE used the program to strengthen its cyber defenses and address the weaknesses it found. The exchange’s parent company is Intercontinental Exchange, which also operates other financial market infrastructure globally.
Disclosure Came During Congressional Testimony
Huizenga asked Martin about the exchange’s efforts to test its systems against coordinated attacks. Martin pointed to Project Glasswing as evidence that the NYSE is actively working to improve its defenses.
She did not provide details on the scale of the vulnerabilities found or the resources used to fix them. Her testimony also did not mention any active threats or breaches, indicating the testing appears to have been proactive rather than a response to an attack.
Cybersecurity has become an increasing concern for financial market operators in recent years, with regulators and lawmakers pressing major exchanges to show they can withstand sophisticated attacks. The NYSE processes millions of trades each day, making it a high-value target for bad actors.
Anthropic has not publicly listed which organizations are participating in Project Glasswing, and the NYSE’s involvement was first disclosed through Martin’s congressional testimony. That limited disclosure leaves the exchange’s use of the program notable mainly as a public example of a major market operator testing AI tools in a defensive setting, rather than as a detailed case study.
Intercontinental Exchange shares fell 1.39% on the day of the testimony.
No further comment from Anthropic or the NYSE was available at the time of publication.
The disclosure was first reported through Martin’s testimony before Congress and circulated in a post from Wall St Engine on X.
NYSE USED ANTHROPIC’S MYTHOS TO FIND CYBER FLAWS NYSE President Lynn Martin says the exchange used Anthropic’s Project Glasswing to identify multiple vulnerabilities in its cybersecurity systems and fix them quickly. Project Glasswing gives vetted partners early access to… pic.twitter.com/6aTnhSymso
— Wall St Engine (@wallstengine) September 2, 2026
