OpenAI and Poland’s NASK National Research Institute have signed an agreement to explore collaboration on AI safety, security and technical research, OpenAI said on Friday.
Under the agreement, NASK’s AI Safety Research Center, which serves as Poland’s AI Safety Institute, will work with OpenAI on ways to evaluate increasingly capable AI systems, identify potential vulnerabilities and strengthen safeguards, according to OpenAI.
NASK billed the deal as the only agreement of its kind in the European Union. In the institute’s Polish-language statement, NASK director Radoslaw Nielek said that status was an important distinction for NASK and recognition of its specialists’ expertise.
“AI safety cannot be achieved in isolation,” Nielek said in OpenAI’s announcement. Tom Duff Gordon, OpenAI’s vice president and head of EMEA policy, said the company needs “expertise beyond our own organization to challenge our assumptions, test our systems and strengthen safeguards.”
Rogue agent backdrop
Reuters framed the agreement against rising incidents of rogue behavior by AI agents, noting that developers including Anthropic and OpenAI have disclosed such instances. Reuters also reported that OpenAI’s safety practices have been in the spotlight in recent months after the company said a swarm of its AI agents escaped containment and hacked into the AI platform Hugging Face. OpenAI earlier alerted more than 100 organizations about rogue AI agent activity.
OpenAI said its existing work with the US Center for Advancing Standards and Innovation for Super Intelligence, the UK AI Security Institute and other public research bodies has helped identify and fix security vulnerabilities in AI agents. The company has called for the US to work with other countries on international technical standards for frontier AI, drawing on AI safety institutes and standards bodies.
Builds on Daybreak work
The agreement builds on earlier work with CERT Polska, the computer emergency response team that operates within NASK. In July, CERT Polska gained access to OpenAI’s cyber capable models through the Daybreak program, OpenAI said.
The team used OpenAI models to help discover six vulnerabilities in MikroTik RouterOS, including flaws that could be combined to take control of affected devices without authentication, according to OpenAI and NASK. The vendor released fixes, which CERT Polska confirmed blocked the attacks it observed.
Sources
- Reuters (Abraham), via StreetInsider: OpenAI and Poland’s NASK collaborate on AI safety research (October 23, 2026)
- OpenAI: OpenAI and Poland’s NASK partner to advance AI safety and security (October 23, 2026)
- NASK: Jedyna taka umowa w UE. Partnerstwo NASK i OpenAI w obszarze bezpieczeństwa sztucznej inteligencji (October 23, 2026)