NASK says the agreement is the only one of its kind in the European Union.
NASK’s AI Safety Research Center, which acts as Poland’s AI Safety Institute, will lead the work. The organizations say they will assess increasingly capable AI systems, identify weaknesses, test safeguards, and study how AI can be used maliciously. NASK also names synthetic-content detection and risk differences across conversation languages as areas of its existing work.
The announcement describes the agreement as a next step after CERT Polska, part of NASK, received access in July through OpenAI’s Daybreak program to advanced cybersecurity models. NASK says CERT Polska found six vulnerabilities in MikroTik RouterOS, some of which could be chained to take control of a device without logging in, and that the vendor released fixes.
The source does not provide the agreement text, a budget, a timetable, model-access details, named research deliverables, or independent evidence of outcomes from the new partnership.