AI · Friday, October 23, 2026
OpenAI and Poland's NASK say they will explore AI safety work
The File desk · Oct. 23, 2026, 7:00 a.m. ET
Partially confirmed1 primary sources
Status
The agreement page is the AI desk's read. openai.com returned an error to a direct fetch on this pass. The page is dated Oct. 23, 2026. The desk places the post at 8:00 p.m. Eastern on Thursday.
OpenAI and NASK, a Polish research institute, said they will explore working together on AI safety. The page is dated Friday, Oct. 23. The AI desk's read of the feed puts it at 8:00 p.m. Eastern on Thursday, from midnight GMT.
The text says the two sides will explore collaboration. It does not list a dollar amount or a product they must deliver.
Radosław Nielek is quoted: AI safety cannot be achieved in isolation.
Why it matters (Estimate). The agreement is a plan to explore work, not a contract with a price. Polish researchers are the named partners, and the open questions are listed with the story. (Estimate.)
OpenAI's account of the page also refers to a July item with CERT Polska about six vulnerabilities in MikroTik RouterOS. That history is OpenAI's telling.
How we know
Notes
- 1.AI desk: page dated October 23, 2026, RSS midnight GMT Friday, which is 8:00 p.m. Eastern Thursday. Quote from Radosław Nielek: AI safety cannot be achieved in isolation.
What is still unknown or disputed
- No dollar figure is in the desk's read.
- This pass did not re-open openai.com.
Primary sources
Every claim in this story is drawn from the documents below. If a fetch failed, that is recorded on the card.
Source 1
OpenAI and NASK
OpenAI, as read by the AI desk · October 23, 2026
The desk read a page dated Oct. 23 saying the parties will explore AI-safety collaboration, with a quote from Radosław Nielek. This pass did not re-open the page.
https://openai.com/index/openai-nask-ai-safety-security-partnership
Ask this story
Answers come only from this story and its sources. Free test: limited questions per day.