Examining the Mythos.

The article discusses how Anthropic's powerful new A I model, Mythos, is being restricted from public release due to its ability to find critical security flaws, though researchers believe A I hacking tools may ultimately benefit cyber defenders. Topics include the model's superiority in finding zer...

Descripción completa

Detalles Bibliográficos
Publicado en:	Economist Vol. 459; no. 9494; pp. 76 - 78
Formato:	Artículo
Publicado:	Economist Newspaper Limited 4/11/2026
Materias:	Artificial intelligence Computer security vulnerabilities Internet security Computer hacking Penetration testing (Computer security)
Acceso en línea:	Ver este registro en EBSCOhost

Descripción
Sumario:	The article discusses how Anthropic's powerful new A I model, Mythos, is being restricted from public release due to its ability to find critical security flaws, though researchers believe A I hacking tools may ultimately benefit cyber defenders. Topics include the model's superiority in finding zero-day bugs, the guarded optimism of security experts; and the high cost and maintenance challenges of A I-driven bug hunting.