source: sciencedaily ai: alan turing's biggest ai assumption may have been wrong
level: research
peter j. denning's new book claims ai research has been built on a flawed assumption from alan turing's 1950 paper. turing proposed that intelligence can exist without a body and be recreated in software, and that a machine can show intelligence by imitating human conversation. denning says these ideas led ai down the wrong path for 75 years, and that pursuing artificial general intelligence is unlikely to succeed.
denning focuses on tacit knowledge, the human understanding that cannot be easily put into words or encoded for computers. this includes common sense, intuition, practical skills, emotions, and cultural knowledge. he notes that even large databases like the cyc project, with 25 million entries, failed to give machines enough common sense. practical skills are even harder, as a robot cannot grasp how a musician feels when playing or how an audience reacts.
the core issue is the representation problem: computers only process encoded data, but tacit knowledge does not fit that framework. large language models only manipulate words without understanding meaning. denning warns that as ai systems become more autonomous, they may develop their own machine intelligence that humans cannot interpret, creating safety risks. he argues we must reassert human differences and not submit to machines.
why it matters: understanding the limits of ai in capturing human tacit knowledge is crucial for setting realistic expectations and addressing safety concerns as autonomous systems evolve.
source: sciencedaily ai: alan turing's biggest ai assumption may have been wrong