One problem with AI that few address is that an LLM model tends to be domain-specific. Humans tend to generalize and use an AI as if it’s a generalist.

#ai #human #failure