OpenAI and Poland’s NASK sign AI safety and security agreement

OpenAI announced on October 23 that OpenAI and Poland’s NASK National Research Institute signed an agreement to explore collaboration on AI safety, security and technical research.

OpenAI and NASK’s AI Safety Research Center plan to examine ways to evaluate increasingly capable AI systems, identify potential vulnerabilities and strengthen safeguards. According to the announcement, NASK’s center serves as the Polish Artificial Intelligence Safety Institute.

The parties said the planned work is intended to deepen their understanding of emerging risks and support a more secure and resilient digital ecosystem. Those are goals, not completed results.

“AI safety cannot be achieved in isolation,” NASK Director Radosław Nielek said in the announcement. Tom Duff Gordon, OpenAI’s vice president and head of EMEA policy, said the agreement would bring Polish research expertise into international work on AI safety.

NASK established its AI Safety Research Center in December 2024, saying it would focus on safe, transparent and explainable AI methods for areas including public administration, network security and cyber protection.

The agreement builds on earlier work between the organizations. NASK said in July that CERT Polska had received access to OpenAI’s GPT-5.5 Cyber through a government trusted-access program. CERT Polska operates within NASK.

OpenAI’s announcement did not disclose funding, commercial terms, a research timetable, named deliverables, model-access commitments or an evaluation protocol for the new agreement.