Correct. I've seen some very expensive experiments with it reduced from "AI" to "machine learning" because the young people couldn't provide good enough input for it to do anything else. They did not train in the field they wanted to apply the computer to, so the computer symptoms were awful and caused decisions that lost money. Pattern recognition still depends on the priorities provided by humans, among other things, and the people behind the wheels of these projects just aren't good at this. Maybe we should be thankful for that