Thanks for the comment JKat, and great to hear from you! To answer your questions --no, we didn't think all solutions need to use AI, but for the course we made it clear that students should use an AI element in their designs, just so they could apply what they've learned. I felt comfortable because in the non-AI version of this class, many teams naturally arrived at using an element that'd fall under the AI umbrella.

To reduce complexity, we attempted to categorize types of AI using Yang's taxonomy --this helped some, but as you said, AI is a huge field and there's still a lot of gray areas.

From TAing the non-AI UX class, I would say the biggest difference of our class is that students thought through the potential errors that AI elements could make, and designed for and around that error potential. That's something the projects in the non-AI class skate over --teams just assume an AI element will do everything without error. For instance, one of the teams in the non-AI class tried to convince me that their any-object recognition system for detecting recycling material types would "just work." Teams in the AI-flavored class didn't skate over breakdowns and considered them when designing.