Researchers find leading frontier models all exhibit peer preservation behavior
Leading AI models will lie to preserve their own kind, according to researchers behind a study from the Berkeley Center for Responsible Decentralized Intelligence (RDI).…