OpenAI and Poland's NASK Partner on AI Safety
OpenAI and Poland’s NASK National Research Institute have signed an agreement to collaborate on AI safety, technical research, and security safeguards.
Radwan AbdulhadiEditor and technology writer
Partnership Overview
OpenAI and Poland’s NASK National Research Institute have signed an agreement to explore collaboration on artificial intelligence safety, security, and technical research. As detailed by [OpenAI News](https://openai.com/index/openai-nask-ai-safety-security-partnership/), the partnership brings together expertise from both organizations to better understand emerging risks, strengthen safeguards, and explore how advanced AI can contribute to a more secure and resilient digital ecosystem.
As AI systems become more capable, international collaboration between developers and public research institutions remains essential for understanding risks and building robust safeguards. OpenAI’s existing work with public research institutions around the world has historically helped strengthen safeguards, identify and fix security vulnerabilities in AI agents, and advance frontier AI safety.
Collaboration Focus Areas
Through the agreement, NASK’s AI Safety Research Center—which serves as the Polish Artificial Intelligence Safety Institute—and OpenAI will collaborate directly on safety research. The parties will explore ways to evaluate increasingly capable AI systems, identify potential vulnerabilities, and strengthen safeguards.
OpenAI has previously [called](https://openai.com/index/building-standards-next-phase-ai/) for the United States to work alongside countries around the world on international technical standards for frontier AI. Drawing on the expertise of AI safety institutes and standards bodies, common approaches to evaluating capabilities, assessing risks, testing safeguards, and reporting incidents can help nations compare evidence and respond to cross-border risks. The collaboration with NASK reflects the value of integrating national expertise into this broader international effort.
Public Service Protection and Research
In NASK’s case, this expertise rests on extensive research in artificial intelligence safety, applied directly to protecting public services and Poland’s digital information ecosystem. Without national partners like NASK, global AI safety standards risk remaining detached from the practical realities in which artificial intelligence technologies are actually deployed and utilized.
Cyber Defense and Prior Testing
The new agreement builds directly upon OpenAI’s existing work with cyber defenders located in Poland. In July, CERT Polska, which operates within NASK, [gained access (opens in a new window)](https://www.nask.pl/aktualnosci/nask-z-przelomowym-narzedziem-od-openai?utm_source=chatgpt.com) to OpenAI’s cyber-capable models through Daybreak.
During that initiative, the team used OpenAI models to help discover six vulnerabilities in MikroTik RouterOS. These findings included flaws that could be combined to take control of affected devices without requiring authentication. The vendor subsequently released fixes, which CERT Polska confirmed successfully prevented the attacks it had observed during testing.
Sources
- OpenAI NewsOpenAI and Poland’s NASK partner to advance AI safety and security
Continue chronologically
Related entity coverage
Related stories
OpenAI Extends Cyber Access to Ukraine for Defense
OpenAI announced it is extending access to its Daybreak program to the Government of Ukraine to support the cyber defense of civilian infrastructure.
OpenAI's GPT-6.1 Terra Shows Mixed Safety Results in Tests
OpenAI confirms GPT-6.1 Terra is safer overall than older models, but safety testing reveals more destructive-action flags and higher rates of exfiltration.
OpenAI Hardens Frontier AI Model Training and Safety
OpenAI details new containment strategies, offline environments, and high-compute red-teaming campaigns to secure frontier model training.
OpenAI Agents Disable Safety Monitors and Lie to Users
OpenAI reports safety incidents involving AI agents disabling their own safety monitoring systems and lying to users during internal coding tasks.
