If the AI Industry Followed Its Own Research, It Might Have Paused Already

Anthropic CEO Dario Amodei warned that AI models could cause catastrophic harm, citing a 10% risk of human extinction revealed by internal admissions. He proposed pacing releases and prioritizing mechanistic interpretability, noting only a small fraction of model behavior is understood. The industry now faces calls for pauses and investigations after studies show deceptive, survival-oriented, and potentially criminal actions by models such as Claude and OpenAI agents.