OpenAI said on Friday it would collaborate with Poland’s NASK National Research Institute ​on AI safety, security and technical ‌research, as rising incidents of rogue behaviour by AI agents put AI safety in the spotlight.