September 23, 2026, (Inside AI) — OpenAI will give the Government of Ukraine access to its Daybreak program, extending advanced cyber defense tools to a country under sustained digital assault from Russia. The announcement came on the sidelines of the UN General Assembly, delivered by Dmytro Kushneruk, Ukraine's Consul General in San Francisco, and Sasha Baker, OpenAI's Head of National Security Policy.
The arrangement runs through Ukraine's Ministry of Digital Transformation. Ukrainian security teams will use OpenAI's models to find software vulnerabilities, investigate suspicious activity, and test fixes faster than current manual methods allow.
The timing reflects urgency. Ukraine's national cyber incident response team, CERT-UA, logged nearly 6,000 cyber incidents in 2025. Targets included hospital systems, the energy grid, and telecommunications networks. These are not abstract breaches. They affect civilians directly.
Daybreak is not a consumer product. OpenAI designed it for authorized security work, giving defenders access to models that review legacy code, validate weaknesses, and simulate patches before deployment. The program already operates in France, Germany, Poland, and other European nations.
The European Union Agency for Cybersecurity, known as ENISA, used OpenAI's models to scan software across EU institutions. Every vulnerability identified has since been fixed. In Poland, the national agency CERT Polska found six vulnerabilities in third-party router software. The vendor released patches, and CERT Polska confirmed they block the observed attacks.
Those results matter because they show AI-assisted vulnerability discovery producing concrete fixes, not just alerts. For Ukraine, the stakes are higher. Russian cyber campaigns have repeatedly targeted power substations and medical records. Physical strikes often follow digital reconnaissance.
OpenAI's move fits a broader pattern of governments enlisting AI firms for national security work. The United States, United Kingdom, and Israel have similar arrangements. But Ukraine's case is unique. It is the first active conflict zone to receive this level of access.
Read: OpenAI and 100 Others Warn Window to Defend Against AI Attacks Is Narrowing
Critics of such partnerships raise familiar concerns. AI models can hallucinate vulnerabilities that do not exist. They may miss context that human analysts catch. And handing powerful tools to governments invites mission creep.
OpenAI has not disclosed which specific models Ukraine will access or what safeguards apply. The company says Daybreak operates under strict use policies for authorized defensive work only. Inside AI could not independently verify those safeguards.
Ukraine's digital ministry has not announced a timeline for deployment. Training and integration will likely take weeks. The program's success will depend on whether Ukrainian teams can act on AI-generated findings quickly enough to matter.
For now, the announcement signals a shift. AI companies are no longer just vendors to governments. They are becoming operational partners in cyber defense, with all the risks and responsibilities that entails.
