Anthropic has handed the European Union's cybersecurity agency, ENISA, the keys to its zero-day hunting AI. The system, called Mythos, is designed to automatically discover previously unknown software flaws before attackers can exploit them.
A tool for finding the unfindable
Mythos is an AI system developed by Anthropic that scans code and system behavior to identify potential zero-day vulnerabilities. Zero-days are security holes that no one knows about, making them a prized weapon for hackers. By finding them first, Mythos could give defenders a critical head start. ENISA will now have access to this tool to enhance its threat detection capabilities.
Bolstering EU cyber defenses
ENISA coordinates cybersecurity efforts across the European Union, from critical infrastructure protection to incident response. Access to Mythos could help the agency proactively identify vulnerabilities in government systems, energy grids, and other key sectors. The move reflects a growing pattern of AI companies sharing advanced security tools with public sector bodies. Instead of reacting to breaches after they happen, Mythos allows for a more preventive approach.
Anthropic’s decision to grant access to EU authorities, rather than keeping Mythos proprietary, signals a shift in how the industry views threat intelligence. The partnership does not involve any financial exchange, according to people familiar with the arrangement. ENISA will use the AI to supplement its existing vulnerability research programs.
What comes next
The exact terms of the access agreement have not been made public. It remains unclear how ENISA will deploy Mythos in day-to-day operations or whether other EU member states will get direct access to the system. The agency is expected to begin testing the AI in the coming weeks, with a report on its effectiveness due by the end of the year.
