Robert long on artificial sentience

Abstract

The LaMDA controversy, where a Google engineer claimed the chatbot LaMDA was sentient, is analyzed through the lens of the philosophy of consciousness and AI safety. The author argues that the discussion conflates several distinct questions about AI, including its understanding of language, intelligence, and sentience. He defines sentience as the capacity to feel pleasure or pain, and consciousness as the ability to have subjective experiences. The author considers various theories of consciousness, such as Predictive Processing, Global Workspace Theory, Integrated Information Theory, and Attention Schema Theory, to assess their potential to explain artificial sentience. He explores the implications of a future filled with digital minds, which may be capable of intense suffering, and highlights the importance of AI alignment in ensuring their benevolent development. The author concludes that although the question of artificial sentience is fascinating, focusing on AI alignment is currently more pressing. – AI-generated abstract.