Harvard Psychologist Pinker Says AI Doom Risk Is Overblown

Harvard psychologist Steven Pinker published an open letter to tech blogger Scott Alexander in Quillette, arguing that the risk of AI destroying humanity has been exaggerated. Alexander had invited him to a public debate on AI existential risk, but Pinker declined, calling such debates "spectator sports."

Pinker said doomsday scenarios like the "paperclip maximizer" confuse intelligence with a desire for domination. He admitted he underestimated how capable large language models would become and did not expect AI companies to release products so recklessly.

The risks Pinker truly takes seriously are bioterrorism, cyberattacks, and unchecked AI agents. He wrote that doom scenarios do not help address these dangers, and that effective action means careful safety engineering based on independent oversight, accountability, and human control.

Alexander argues there is a 20% chance AI causes human extinction and calls for better AI safety research and international agreements to slow AI development.