Steven Pinker on AI Doom and Consciousness

Steven Pinker thinks the case for AI doom rests on confused assumptions about intelligence and motivation.In September, the rationalist and blogging legend Scott Alexander challenged him to a public debate on the topic. Pinker declined, setting out his objections in an open letter instead, and Alexander has since replied at length.In this episode, Henry Shevlin and I talk to Pinker about how his views on AI connect to his broader ideas about intelligence, motivation, evolution, and progress. We also put several objections to him (including some that feature in Alexander’s reply) and ask whether the rapid progress in AI over the past few years has changed his mind about anything.It has, at least when it comes to neural networks.Along the way, we discuss instrumental convergence, evolutionary analogies for AI, whether AI systems could be conscious, why he would not sacrifice any human welfare for the sake of AI welfare, and why so many intellectuals are pessimistic about technology. At the end, we ask him what would have to happen for him to concede that Alexander and others were right. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.conspicuouscognition.com/subscribe