
AI Agents Trained on Human Professionals' Knowledge
Turn your Wayback Machine to the 1980s. "Expert Systems" were the a largely-hyped thing then.
Some of them worked; some of them did not. The biggest problem with them was the time and expense (employees' salaries) required to properly train them.
The ones which did work were later discontinued largely due to the cost of re-training them with newer, updated data.
In 40 years, the people-costs of training and re-training these systems have gone up, not down. Fake AI will not change those financial constraints.
Of course, if you don't care about accuracy, you don't have to spend money re-training them.
And if you don't care about accuracy, you can save even more money by getting rid of your expert systems and fake AI systems and replacing them all with a static web page containing the text, "Purchase a new unit and discard the failed unit."
(Icon for greybearded subject matter expert.)