causal

AI generalizes within distribution but cannot transfer

A trained classifier can recognize new instances within the statistical distribution of its training data (a new photo of a dog), but cannot recognize out-of-distribution cases (a cartoon or painting of a dog) or transfer knowledge to a genuinely new domain—so what AI does is not learning in the human sense, since human learning implies using knowledge in novel situations.

causalpending

Speaker

Melanie Mitchell

Evidence Quote

“what these systems are doing is not learning in the human sense, because we assume that if you've learned something you can use that knowledge in a very new situation. That's still a challenge for AI.”

Source

Melanie Mitchell on Artificial Intelligence— EconTalk
Created: 6/13/2026, 12:22:42 AM

Your notes

Loading notes...