AI's Wrong Path: Turing's Mistake Exposed

Computer scientist Peter J. Denning asserts in "Turing's Mistake" that Alan Turing's 1950 propositions about artificial intelligence may have misdirected research for 75 years. Denning contends that AI's pursuit of intelligence independent of a body and the Turing test's conversational imitation criteria have led to current challenges.

Central to Denning's argument is the concept of tacit knowledge, which includes common sense, emotions, and practical skills that are difficult to encode for machines. He notes that even extensive projects like Cyc, aimed at cataloging common sense, have fallen short of creating true expert systems.

Denning emphasizes that the "representation problem" prevents machines from grasping the embodied nature of human knowledge. He explains that Large Language Models like ChatGPT manipulate symbols but lack genuine understanding of meaning, as the connection between words and tacit knowledge remains a mystery.

Furthermore, Denning argues that intelligence is deeply intertwined with context and culture, elements that are inherently difficult for machines to replicate. He predicts that the divergence in tacit knowledge between humans and machines could lead to significant AI safety concerns, with machine intelligence posing a greater threat than superintelligence.

Denning calls for a reassertion of humanity, recognizing the unique differences between humans and machines, rather than striving to replicate human intelligence. He stresses the importance of learning to coexist safely with AI, acknowledging that the familiar cultural landscape is evolving.