Brilliant post on a crucial topic: ‘Intelligence’ is a root concept, and as such, it informs all other concepts. If we counterfeit or distort it by pretending that machines ‘can be intelligent’ (something that is formally impossible unless we so warp the definition of intelligence that objects can qualify) we are damaging our own root definitions in ways that could prove unrecoverable over time.

It is my positions that we must be extremely careful about how we use important root concepts… and protect them from falsification. Computer-assisted problem solving is a real phenomenon, ‘artificial intelligence’ is dangerous nonsense. A related topic is ‘machine learning’; machines cannot properly ‘learn’, what they can do is form models from data. I would suggest that we use terms like ‘computational heuristics’ to describe these technologies for the sake of preserving the meanings of learning and intelligence, and, specifically, limiting them to living systems and human beings.