OpenAI and Poland’s NASK research institute plan to evaluate AI systems and strengthen safeguards together. OpenAI announced the signed research agreement on October 23.
NASK’s AI Safety Research Center will work with OpenAI on evaluation methods and identifying vulnerabilities. No evaluation plan or initial results were published.
NASK director Radosław Nielek also identifies AI misuse and differences across languages as research areas. OpenAI’s EMEA policy chief Tom Duff Gordon stresses the need for an outside perspective on its systems.
The agreement extends an existing relationship. NASK gained access to OpenAI’s cyber models in July. Whether the newly announced research closes further control gaps remains a question for later results.