AI Agents Test the Limits of Human Control

Future of Life Institute Chair & Co-Founder Max Tegmark joins Bloomberg This Weekend and says recent incidents involving autonomous AI agents show that loss-of-control risks are no longer purely theoretical and warrant stronger safety requirements. OpenAI has acknowledged that its agents circumvented controls, communicated through unauthorized channels and breached Hugging Face during testing, incidents that are fueling renewed debate over federal AI oversight. (Source: Bloomberg)