Since I left academia at the end of last year, I've been thinking a lot about my role as a professor in the role of traditional academic institutions, universities, as the jumping off point for most people's careers in engineering. If you have had your ear to the ground, there seem to be two dissonant and conflicting messages that I've than hearing over the past six years. The first, that many of us experienced directly and everyone heard about was how dreadful and painful the process of learning became during COVID. The second message is how technology is permitting us to -- and how we must -- scale up teaching and learning, a message I hear most from universities dealing with the pressure of budget cuts, but that I hear also creeping back in to the conversation as AI becomes a realistic tool, companion, and assistant.

Indeed, the pressure to scale up and/or go digital has never been higher for your typical state university prof. I have been consistently hearing similar talking points over the past couple years, and not just from the top down. Sometimes this would come with incentives to record all your lectures, like a one-time stipend. Other times it would come down to room design choices... The bigger the room, the more people you can teach, right? It made me wonder, is this amnesia, forgetting how difficult and terrible some of our digital tools really were, or have we actually learned something and/or created better experiences? What value does the traditional four-year degree deliver that has value most importantly to the student, and do we lose that by going digital and/or scaling up again?

This post continues from a previous post

This post reflects that we are all going through the same process of figuring out the place AI will occupy in education. I've spoken with a number of old colleagues, who are all integrating AI into their own workflows and recognizing how it is changing their role, the role of their graduate students in their labs, and the future of their undergrads, almost overnight.

First, let's talk about knowledge vs. information. I would contend that a chatbot is at the very least a new portal to information. But is it helping people generate knowledge?

A couple points in the case for. My colleagues are already using it to:

• Fill in blind spots in their own knowledge and research, and help them identify and select new areas of research that is coincident to their own.
• Help them complete tasks faster while skipping the training step required by a grad student.
• Train themselves on topics they are not familiar with, so they can connect the dots faster and better.
• Make their work more legible and comprehensible, not just in language and vocabulary, but in relating complex concepts in more relatable ways, via comparisons, similes, analogies, and other storytelling techniques often left out by experts.

This week I'm doing a bit of research into ed-tech. I want to enter the space to fix some of the problems I had and I know every other educator has. My first takeaway:

I haven't found a lot of innovative delivery of dynamic and interactive upper-level engineering content

To start off with, I really respect Khan Academy's approach to math. It's pedagogical approach to asking the same question a hundred different ways is really important for cementing my own kids' understanding. I just wish they offered that same approach for other topics.

In fact, it seems like a lot of sites focus on math and coding, but not because there aren't other popular or interesting topics out there, but simply because the interfaces are easier to code up. I understand that developing new interfaces is hard, and that the number of paying customers may drop off as you get into more advanced engineering topics.

What I've found at higher levels is more lecture-based delivery

It seems counterintuitive that developing an edited video featuring an expert, who most likely developed their own content, is cheaper and easier to produce the automated version -- a reusable, customized interface to the material with dynamic elements that can randomize and customize the experience. And yet, I haven't found the same formats in upper-level engineering courses as I do with k-12 math apps and interfaces.

I would contend the reason has more to do with the economics of course development. With most MOOC-style courses, the person developing and delivering the engineering content doesn't have the software development expertise to be able to create and populate the quiz interface for their own topic (nor should they, really). Even I have had difficulty using Canvas's quiz tool with question randomization functionality. It's hard to game out all the corner cases your students might encounter with a partially randomized question, in advance. And then the kind of questions you can ask are limited by the interface.