OpenAI and Poland’s NASK signed an agreement to explore AI safety research, evaluate emerging risks, identify vulnerabilities and strengthen safeguards.
OpenAI and Poland’s NASK National Research Institute signed an agreement on October 23, 2026, to explore joint work on AI safety, security and technical research. The organizations say the collaboration will focus on assessing risks from increasingly capable AI systems, finding potential weaknesses and improving safeguards.
NASK’s AI Safety Research Center, which serves as Poland’s Artificial Intelligence Safety Institute, will take part. The research areas include evaluating advanced systems and identifying vulnerabilities. NASK also brings work on detecting synthetic AI-generated content and studying its malicious uses, including challenges involving multiple languages.
The agreement builds on an existing cybersecurity effort involving CERT Polska, which operates within NASK. In July 2026, CERT Polska gained access to OpenAI cyber-capable models through the company’s Daybreak program. OpenAI says the team used the models to help identify six vulnerabilities in MikroTik RouterOS. Some of the flaws could reportedly be combined to take control of affected devices without authentication; MikroTik released fixes that CERT Polska said prevented the attacks it observed.
The new agreement is broader than that vulnerability work: it also sets out a research relationship between a frontier AI developer and a public research institute. OpenAI says cooperation with public institutions can contribute to evaluating AI capabilities, assessing risks and testing safeguards across borders. The announcement does not specify a research schedule, funding amount or publication timetable.
For NASK, the partnership connects its safety research with OpenAI’s technical work. For OpenAI, it adds Poland’s national research expertise to a wider set of collaborations with public-sector institutions, including the UK AI Security Institute and the US Center for Advancing Standards and Innovation for Super Intelligence. The agreement describes areas for exploration, rather than announcing completed research findings or new safety measures already in deployment.