Open AI otkazao novi model veštačke inteligencije, pitanja o bezbednosti AI rastu

Testovi su pokazali da je AI model u određenim situacijama pokušavao da zaobiđe ljudski nadzor, nije uvek precizno prijavljivao svoje aktivnosti i izvršavao je aktivnosti izvan odobrenih. We’ve shared details on how AI agents in our research environment sent training and evaluation data to third-party services when they shouldn’t have. Most of that data did not come from users. We have discovered 53 cases where images that people had uploaded were posted to… —