The current hype and the massive investments are about generative AI, not the actually-useful-for-humanity applications.

Image recognition, medical research etc. are not drives the current market. It's about offering a service that the broad masses use continuously. Otherwise these investments don't make sense.