I worked with AI back in the '80s, and back then several start-ups were peddling "Inference engines" for $500 or so a copy -- software that you could feed "rules" and it would give you answers.  They didn't really work -- demonstrations looked great so long as the domain was very limited and artificial -- and they soon disappeared.

We still have materialist-type neurologists insisting computers are capable of great feats of logic, and that is true, and they are wonderful for mass data storage and sorting and fact calculation and modeling, but they don't "experience" the world the way they do and everything is of a reflexive nature, without real sentience.  The problem is we just don't know how the human brain (and many animal brains) produces sentience, and until we do we will not be able to do it in machines.

Of course the efforts should still be carried out.  Maybe they will succeed and thereby we will better understand our brains, but I'm not optimistic and think most of it is hype to get funding or tenure.