Maybe the ebook will be of interest for you, AI guys

Learning how to apply an ANN to solving a particular task would be a financially lucrative move for anyone interested the analysis of data but AI is more of a generic term covering many techniques although it originally meant duplicating the way humans think.

I have been looking at the deep learning examples used with the Jetson Nano but the real question as regards a brain for robot is what it can't do compared with us.  So it can say with 90% accuracy there is a banana in the image but does it tell the machine its orientation, size and how to peel it? Although we do learn some things through many examples (although often a lot less than required by an ANN)  we  can also use that learning to learn more things with only one example.  If I showed you a tropical fruit you had never seen before you would immediately see it as a particular object and be able to recognize it later after seeing it once.