Anthropic restringe el acceso a la IA Claude Mythos a través del Proyecto Glasswing ante temores de hackeo

Tras la revelación de la semana pasada que provocó alarmas mundiales, Anthropic ha restringido su potente IA Mythos —capaz de encontrar vulnerabilidades de ciberseguridad— a empresas seleccionadas bajo el Proyecto Glasswing, incluidas Amazon Web Services, Apple y Google, después de que una filtración accidental generara preocupaciones sobre la seguridad nacional.

Mythos, que surgió públicamente el mes pasado a través de contenido web sin asegurar y acceso a foros privados, puede identificar miles de fallos de gravedad alta y crítica en software y sistemas operativos. Anthropic citó riesgos para las economías, la seguridad pública y la seguridad nacional, optando por un acceso controlado en lugar de un lanzamiento completo. Bajo el Proyecto Glasswing, empresas como Amazon Web Services, Apple, Google, JPMorgan Chase, Microsoft y NVIDIA la están probando para detectar errores en sus sistemas. Los expertos en seguridad están divididos: Davi Ottenheimer la calificó como una herramienta legítima presentada como una amenaza; Kevin Curran, de la Universidad del Úlster, cuestionó que la IA supere a los hackers humanos; Bobby Holley, de Firefox, señaló que encontró 271 vulnerabilidades en el navegador, todas ellas detectables por humanos de élite. El AI Security Institute consideró que solo es efectiva contra sistemas pequeños y débilmente defendidos, aunque avanza rápidamente. Alan Woodward, de la Universidad de Surrey, elogió su velocidad para los atacantes, pero señaló que las defensas siguen siendo viables. Expertos como Holley y Curran ven un potencial proactivo para los fabricantes de software a medida que este tipo de herramientas proliferan.

Artículos relacionados

Illustration of Anthropic restricting Claude Mythos AI and launching Project Glasswing consortium with tech giants to address cybersecurity vulnerabilities.
Imagen generada por IA

Anthropic restricts Claude Mythos AI release and launches Project Glasswing over cybersecurity risks

Reportado por IA Imagen generada por IA

Anthropic has limited access to its Claude Mythos Preview AI model due to its superior ability to detect and exploit software vulnerabilities, while launching Project Glasswing—a consortium with over 45 tech firms including Apple, Google, and Microsoft—to collaboratively patch flaws and bolster defenses. The announcement follows recent data leaks at the firm.

Anthropic has released a new cyber-focused AI model called Mythos, capable of detecting software flaws faster than humans and generating exploits. The model has raised alarms among governments and companies for potentially turbocharging hacking by exposing vulnerabilities quicker than they can be patched. Officials worldwide are scrambling to assess the risks.

Reportado por IA

In the wake of Anthropic's unveiling of its powerful Claude Mythos AI—capable of detecting and exploiting software vulnerabilities—the US Treasury Secretary has convened top bank executives to highlight escalating AI-driven cyber threats. The move underscores growing concerns as the AI is restricted to a tech coalition via Project Glasswing.

Anthropic's CEO Dario Amodei stated that the company will not comply with the Pentagon's request to remove safeguards from its AI models, despite threats of exclusion from defense systems. The dispute centers on preventing the AI's use in autonomous weapons and domestic surveillance. The firm, which has a $200 million contract with the Department of Defense, emphasizes its commitment to ethical AI use.

Reportado por IA

Anthropic has launched Claude Opus 4.7, a new AI model designed to assist developers with complex coding tasks. The company emphasized its improved instruction-following and memory capabilities. This release follows the earlier announcement of the more advanced Claude Mythos Preview.

Global investors are questioning the returns on massive tech spending in artificial intelligence. Christopher Wood, from Jefferies, identifies Anthropic as a standout in the evolving AI landscape. The AI boom has boosted US equities, but concerns grow over its sustainability.

Reportado por IA

Anthropic has restricted unlimited access to its Claude AI models through third-party agents like OpenClaw, requiring heavy users to pay extra via API keys or usage bundles starting April 4, 2026. The policy shift, announced over the weekend, addresses severe system strain from high-volume agent tools previously covered under $20 monthly subscriptions.

 

 

 

Este sitio web utiliza cookies

Utilizamos cookies para análisis con el fin de mejorar nuestro sitio. Lee nuestra política de privacidad para más información.
Rechazar