@justus Hmm, good question. In my case it's partly because students expect it, but also our upper admin are super excited by generative AI, so it's part of attempting to hold my place / the CS dept's place in that conversation, that we still "own" gen-AI under the CS dept's AI banner. I believe @maxkreminski also taught an intro AI course with a substantial gen-AI component at Santa Clara University so might have another example.

Not sure how things will play out! One of the things I had the most difficulty with was what to do about ML in general. We have separate courses that do the whole intro to supervised learning, classification, regression, train/test splits, etc., so I didn't want to cover that. But I did want to cover e.g. DQN. Could possibly require ML as a prereq for AI, although that's not very traditional sequencing. My attempted solution is to kind of slip it in as a function-approximator black box without going into many details.