Polish Research Institute NASK and OpenAI reach agreement on AI-based workplace safety – Unite.AI

OpenAI and the Polish National Research Institute NASK have signed an agreement to explore collaboration on safety, security and technical research on artificial intelligence, OpenAI announced on October 23, 2026.

OpenAI said the partnership will bring together the expertise of both organizations to better understand emerging risks, strengthen safeguards and explore how advanced artificial intelligence can contribute to a more secure and resilient digital ecosystem. Under the agreement, NASK’s AI Security Research Center, which serves as the Polish Institute for AI Security, and OpenAI will collaborate on AI security research, exploring ways to evaluate increasingly capable AI systems, identify potential vulnerabilities and strengthen safeguards.

OpenAI said existing work with the U.S. Center for Advancing Standards and Innovation for Super Intelligence, the U.K.’s AI Security Institute and other public research institutes has already helped strengthen safeguards, identify and fix security vulnerabilities in AI agents, and advance frontier AI security.

Statements from NASK and OpenAI leadership

NASK Director Radosław Nielek said that collaboration with OpenAI is a natural step for the institute and its Artificial Intelligence Security Research Center. He said the center’s work ranges from developing safeguards against emerging risks, to detecting synthetic AI-generated content, and understanding both the harmful uses of AI and the multilingual aspects of these challenges. AI safety cannot be achieved in isolation, he said, and only cooperation between public AI safety institutions and frontier AI developers can make AI truly reliable and safe to use. By combining NASK’s research expertise with OpenAI’s technical capabilities, he said, the two organizations can better anticipate risks, strengthen safeguards and build public trust in technologies.

Tom Duff Gordon, vice president and head of EMEA policy at OpenAI, said Poland is home to exceptional scientists and technical talent, many of whom have helped lead OpenAI’s research. He said that as AI becomes more capable it is crucial to draw on expertise outside OpenAI’s own organization to challenge its assumptions, test its systems and strengthen safeguards, and that the deal with NASK opens the door to do this together, while bringing Poland’s research expertise into the international effort to make AI safer.

A broader push for international AI standards

OpenAI said the NASK collaboration reflects the value of bringing national expertise to its broader call for international technical standards for frontier AI. The company urged the United States to work with countries around the world on such standards, drawing on AI security institutes and standards bodies, and said common approaches to capability assessment, risk assessment, safeguards testing and incident reporting can help countries compare evidence and respond to risks that cross borders. Without national partners like NASK, OpenAI said, global AI safety standards could remain detached from the realities where AI is actually used.

In a September 21, 2026 post, OpenAI said the United States should lead an effort to develop global technical standards for frontier AI, including recursive self-improvement, a process in which AI systems increasingly guide the development of subsequent generations of AI. The post mentions three challenges that such standards would address: fragmentation, where different nations’ assessments, reporting requirements and incident definitions could conflict; collective action, in which nations acting independently can produce results that no nation wants; and uneven capacity, because frontier development activity and related expertise are not distributed equally across the world.

The post pointed to AI safety institutes already established in Australia, Canada, Germany, France, Kenya, Japan, Korea, Singapore, India and the United Kingdom as a network that could facilitate standards-setting.

He proposed standards for evaluating AI advances relevant to recursive self-improvement and the amount of autonomous research within an AI company; for human oversight of automated AI research, including which automated research processes should trigger immediate human review; and for incident classification, monitoring, reporting and response, such as common severity levels and reporting thresholds. OpenAI said these technical standards would not be licensing, mandatory pre-release review or approval requirements for AI models, and that national governments would decide whether and how to incorporate them into their own legal systems.

Early work on cybersecurity in Poland

The agreement also builds on OpenAI’s existing work with cyber defenders in Poland. In July 2026, CERT Polska, operating within NASK, gained access to OpenAI’s cyber-capable models via Daybreak. OpenAI said the CERT Polska team used its models to help uncover six vulnerabilities in MikroTik RouterOS, including flaws that could be combined to take control of affected devices without authentication. The vendor released fixes which, according to OpenAI reporting, CERT Polska confirmed prevented the observed attacks.

NASK’s July 2, 2026 announcement states that the institute has gained access to GPT-5.5 Cyber, a specialized artificial intelligence model created by OpenAI to detect and analyze cyber threats, through the Government Trusted Access for Cyber ​​program for government units. NASK called Poland the third European country, and one of the few in the world, to receive such access. He described Daybreak as an initiative to keep defenders one step ahead of criminals in using artificial intelligence for cybersecurity, with GPT-5.5 Cyber ​​applied to malware analysis, digital forensics and vulnerability discovery to accelerate the entire cycle of vulnerability detection and remediation.

Nielek said at the time that AI tools remain tools that only work properly in human hands, with a person always at the end of the analysis performed by a model and verifying what the tool has done. CERT Polska head Marcin Dudek said the team’s experts need access to such tools to efficiently detect and repair vulnerabilities before attackers do, and added that CERT Polska already shares its insights from using large language models on its website and plans to continue.