
Artex-linked attacks can move from targeting to compromise in under three hours, according to the FBI.
OpenAI is expanding access to Daybreak, its cybersecurity tool, for NATO and several European governments, with early use cases focused on software audits, vulnerability discovery, and cyber defense support.

OpenAI is giving NATO and several European governments access to Daybreak, its cybersecurity tool for cyber defense teams. The tool lets users apply OpenAI models to audit software, find vulnerabilities, and patch security gaps. France, Germany, Poland, and Ukraine are among the governments named in the report.
The move positions AI models as operational support for public-sector cyber defense rather than just analysis tools. For readers, the key takeaway is that AI-assisted vulnerability discovery is moving into government and alliance-level security workflows.
The NATO Communications and Information Agency says it is already making heavy use of the models behind Daybreak. NCIA chief Dylan Browne described the capability as a generational leap in cyber defense.
During NATO’s Crossed Swords 2026 exercise, participants used AI agents for cyber operations under human oversight for the first time, according to the article. That detail matters because it frames Daybreak-style systems as tools for supervised operations, not fully autonomous cyber activity.
ENISA, the EU’s cybersecurity agency, has already used the models to find four vulnerabilities in software used by EU institutions. A joint report with CERT-EU examined 27 software projects, with 95 percent of 88 reported issues turning out to be real bugs.
Poland’s CERT Polska also found six vulnerabilities in MikroTik routers that were being actively exploited. These examples show how government teams are applying AI to practical vulnerability discovery and validation, not just research demonstrations.
Daybreak’s rollout suggests AI systems are becoming part of the defensive toolkit for governments facing software supply-chain risks and active exploitation. The strongest near-term use case in the provided report is targeted support for finding and patching weaknesses faster.
The human-oversight detail remains important: the article describes AI agents being used under supervision during NATO exercises. That points to a model where cyber teams use AI to accelerate analysis while keeping people responsible for decisions and operations.

Artex-linked attacks can move from targeting to compromise in under three hours, according to the FBI.

The lawsuit targets alleged rogue agent incidents and could test legal accountability for AI labs.

The AI cybersecurity startup raised $125M in a NEA-led Series A.

AI projects are shutting down, pivoting, or getting folded into bigger platforms.