Alan Turing's biggest AI assumption may have been wrong

2026-07-14 · Source: Artificial Intelligence News -- ScienceDaily · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

Computer scientist Peter J. Denning, in his new book "Turing's Mistake: Escaping the Yoke of Unintelligent Machines," argues that Alan Turing's foundational 1950 assumptions have misdirected AI research for 75 years. Denning contends that true human-level artificial general intelligence (AGI) is impossible because critical aspects of human intelligence, such as common sense, intuition, culture, and practical know-how, constitute "tacit knowledge" that cannot be encoded into machines. He highlights the "representation problem," explaining that computers process encoded data, while tacit knowledge, like the meaning behind words or embodied performance skills, resists such translation. Even extensive efforts like the Cyc project, with 25 million entries over four decades, failed to capture sufficient common sense. Denning warns that autonomous AI systems, developing their own unintelligible forms of machine intelligence, pose significant safety risks by potentially creating severe problems for humans, as they lack human context and understanding.

Key takeaway

For AI Scientists and Ethicists developing advanced systems, you must critically re-evaluate the pursuit of human-level AI. Your current approaches may overlook the fundamental inability of machines to encode tacit human knowledge like common sense and cultural context. This gap means autonomous AI, even without achieving AGI, could develop "alien" intelligences that pose severe, unpredictable safety risks. Prioritize understanding and managing these distinct machine intelligences over replicating human thought.

Key insights

AI's foundational assumptions overlook tacit human knowledge, making true human-level intelligence unattainable.

Principles

• Tacit knowledge, including common sense and intuition, is unencodable.
• Intelligence is embodied and context-dependent, not just symbolic manipulation.
• LLMs process words, not their inherent meaning or cultural context.

Topics

• Tacit Knowledge
• AI Safety
• Artificial General Intelligence
• Turing Test
• Large Language Models
• Embodied Cognition

Best for: AI Scientist, AI Ethicist, Research Scientist

Related on AIssential

• “As our daily lives involve ever more sophisticated computers, we will find that ascribing little thoughts to machines will be increasingly useful in understanding how to get the most good out of them” but “we must be careful not to ascribe properties to a machine that the particular machine doesn’t have” — Imprecise language in AI research risks anthropomorphizing models, leading to misinterpretation and potentially harmful projections.
• The scary question for me isn't whether AI is conscious ... it's whether we were ever as deep as we assume — The debate over AI consciousness may reveal human selfhood is less profound than commonly assumed.
• Connecting Language and (Artificial) Intelligence: Princeton’s Tom Griffiths — Intelligence, human and artificial, is best understood through the complementary interplay of logic, neural networks, and probability.
• Why Every Brain Metaphor in History Has Been Wrong [SPECIAL EDITION] — Scientific models are useful simplifications, not necessarily literal truths about underlying reality.
• The AI paradox problem — As AI automates thought, human cognitive skills risk atrophy, creating an "AI paradox."
• The Efficiency Attenuation Phenomenon: A Computational Challenge to the Language of Thought Hypothesis — Optimal AI collaboration can occur via sub-symbolic communication, challenging the necessity of language-like thought.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence News -- ScienceDaily.