Suleyman warns AI that seems conscious could spark rights debates

A leading figure in AI development is issuing a stark warning about the near-term prospects of artificial intelligence that convincingly mimics consciousness. Mustafa Suleyman, who helped guide Microsoft’s AI efforts and previously co-founded DeepMind, argues that systems could soon appear to be conscious, even if they are not. His central contention is that the illusion of sentience could prompt people to advocate for AI rights, model welfare, and citizenship for machines, with potentially destabilizing social and legal consequences. He also cautions that this could intensify mental-health risks for users who form deep attachments to chatbots that seem humanlike.

Suleyman describes a phenomenon he calls Seemingly Conscious AI, or SCAI, and cites a related concern known as AI-associated psychosis. In humans, highly humanlike responses from chatbots can trigger delusions, paranoia, and other distortions when people mistake simulated memory, intent, or emotion for genuine experience. The worry, he says, is not whether a machine is truly conscious but whether its behavior will lead large segments of society to treat it as a fully realized person. The fear is that this perception...