Bank of Spain urges AI cyber defenses, warns advanced models could amplify financial-sector attacks

The Bank of Spain on May 14, 2026, called for stronger international coordination and broader access to defensive artificial intelligence tools to counter cybersecurity risks from advanced vulnerability-detection models. In its semiannual financial stability report, the central bank said tools such as Anthropic’s Mythos could shorten the time between identifying software flaws and exploiting them. Under adverse scenarios, it warned, such models could create systemic risks by enabling more synchronized cyberattacks across the global financial sector and wider economy. The bank cited Anthropic’s limited-access Glasswing program as an example of protective AI technology and said resilience will depend on how quickly institutions patch vulnerabilities and anticipate emerging threats. Financial Stability Director Daniel Perez Cid said the authority has discussed risks tied to new AI models with Spanish lenders, without giving details.