Steven Pinker rejects AI doom rhetoric and calls for a calmer safety engineering approach

Harvard psychologist Steven Pinker said fears that artificial intelligence could wipe out humanity are overstated, pushing back against extinction-focused narratives around the technology. According to The Decoder, Pinker also declined a public debate with blogger Scott Alexander, describing that kind of event as a form of "spectator sport." Alexander has put the probability of AI destroying humanity at 20%. Pinker instead argued for a more measured AI safety framework built around independent oversight, accountability, and human control. The remarks reflect a clear split between catastrophe-centered arguments and a governance-oriented approach that treats AI risk as an engineering and oversight problem rather than an inevitable path to human extinction.