EVERY SIGNAL, SORTED.

Poland, Central Europe, NASK, OpenAI, AI Safety Institutes, Model Evaluation, Multilingual Risk, Synthetic Content, Cybersecurity, Public Research and Accountability

Poland and OpenAI formed a safety partnership. Publish the research queue

NASK and OpenAI signed an agreement to explore model evaluation, vulnerabilities, safeguards, synthetic content, malicious use and multilingual risk. The announcement names important territory, but the useful public artifact will be a dated queue of questions, methods, deliverables and findings.