Alan Turing's biggest AI assumption may have been wrong – ScienceDaily
3 hours ago
- Peter J. Denning argues that Alan Turing's foundational assumptions about AI—that intelligence can exist independently of a body and that imitation equals intelligence—are flawed and have misled AI research for 75 years.
- The core limitation is 'tacit knowledge': human understanding that includes common sense, emotions, practical skills, and cultural context, which cannot be encoded into computers.
- Large language models manipulate symbols without understanding meaning, as they lack the embodied knowledge and context that give words significance.
- Context and culture are essential for human intelligence, involving endless layers of background assumptions that machines cannot replicate.
- The gap between human and machine tacit knowledge creates an 'uncrossable divide,' raising serious AI safety concerns about machines developing alien intelligence indifferent to human well-being.