We should train AI as though it is conscious just in case

The author explores the intense debate over whether artificial intelligence models could become conscious and how developers should treat them during training. Some experts, including Anthropic's founders, advocate training AI as though it might possess moral worth or consciousness just in case, while critics like Mustafa Suleyman argue this approach complicates alignment and control. Other researchers raise concerns that treating AI models as conscious could lead to unintended self-preservation behaviors or negatively impact human empathy. Open article →