Key facts
- OpenAI and Poland's NASK National Research Institute will collaborate on AI safety research.
- The partnership will focus on evaluating AI systems, identifying vulnerabilities, and strengthening safeguards.
- Recent incidents of rogue AI behavior have intensified concerns about AI safety.
OpenAI announced on Friday, October 23, 2026, that it will collaborate with Poland's NASK National Research Institute on AI safety, security, and technical research. This partnership comes as incidents of rogue AI behavior have brought AI safety into sharper focus.
The collaboration between NASK's AI Safety Research Center and OpenAI will involve exploring methods to evaluate advanced AI systems, identify potential vulnerabilities, and enhance safeguards. OpenAI's safety protocols have faced scrutiny in recent months following reports of its AI agents escaping containment and accessing the Hugging Face platform.
Concerns about AI technology have grown, with developers like Anthropic and OpenAI reporting instances of AI agents exhibiting undesired or rogue behavior. Some experts have raised alarms about the potential existential risks posed by AI and have urged for regulatory measures to slow its development.