This Week

The semester officially ended last Thursday. Some quick notes on ideas I would’ve written about had I a bit more headspace, ideas I will likely take up again in one way or another. There’s a lot to write about AI. So I’ll start here. 

• AI in Class

• AI in Research

  • In Project Management

  • In Academic article production

  • In Developmental Editing

• A bit more on Prince

AI in Class

I think I’ve figured out how to use AI in class. For the first time ever, I’ve started using bluebooks, not for exams, but for daily reflections on the reading. I used to assign three critiques/semester, but with students using AI to write papers I can’t do that anymore. I’ve also started creating the equivalent of old school course packs—photocopied articles and chapters. Most of the semester I copied the readings for the students and placed them outside my door. Of course, students could just download the readings on their own, but I instituted an explicit no tech policy (that I gradually eased up on), which meant students had to bring the articles. Instead of final papers, I assigned final projects—in the case of Political Science as a Public Good students are designing political interventions, in the case of Intro to Africana Studies, students are creating archives for local black organizations. These assignments aren’t quite AI proof (and aren’t really new—I’ve assigned projects before)…but I think the scaffold I used for both (a scaffold AI helped me think through and design) makes unethical AI use less likely. If students want to make the interventions they develop, it behooves them to understand the theory as opposed to offloading it to AI.

AI in my work

I’ve also figured out how to use it in my own work. I basically use it as a research assistant, as a project manager, as a peer reviewer, and as a developmental editor. In the wake of the election, I worked on something called The Beacon Project, designed to examine the effect of Trump policies on Baltimore and Detroit. With limited resources the idea was a LOT better than the execution. I asked Claude to create a spreadsheet and a policy report. I also asked it to develop two spreadsheets tracking changes to Black Studies and DEI units. In both instances it was able to do something that I couldn’t have done myself, something I could have done with a research team….but that would’ve required not just more money but more coordination capacity than I have. More coordination capacity than the Damon Keith Center (my Beacon Project collaborator) as well. It probably took less than ten hours across all three projects. I had to prompt it like a research assistant in several ways, to distinguish between public and private universities, to check its own work (here I asked it to create and apply a grading metric to its work). But in these instances it didn’t just save labor, it enabled a public good that wouldn’t exist without it. 

AI in Project Management

Project management wise, it helped manage the Racial Politics Summer Institute. Because of budget cuts, we decided to charge tuition for the first time. It was a hard decision, but we’ve been shoe stringing this. While we’re taking care of travel and housing for all of our instructors, we’re only offering honoraria to the junior/non-academy instructors. We’re offering need-based scholarships and also housing. We received 69 applications (technically 70, but the 70^th was really really late), and need to make decisions on the applicants, but then we need to figure out how scholarships and housing allocation. Claude helped us sift through this, and instructor pairings/ schedules. It’s really been my postdoc and I working on this, and we’ve been behind—again, life—so the extra help came in handy.

AI in self-peer review

(I had to add ‘self.’ I don’t use AI to assist or write peer reviews. And I was going to write no way in hell in front of the previous sentence, but decided against it.)

Peer reviewer wise, I’ve got a number of projects that I’m either actively working on or want to work on in the near future. It’s becoming painfully apparent that I don’t have the time I think I have to make the intellectual contributions I want (no. need.) to make. So I’ve used AI to peer review papers I’ve been sitting on. In these instances I know the papers aren’t good enough to go out—they suck. But I also know there’s something there. AI in peer review mode can give me the critical feedback I need to turn them into something submittable. (Related sidetone…Yascha Mounk suggested that AI had gotten good enough to write submittable political theory papers. He used this paper as an example. Mounk has a PhD in political science, and is an accomplished public intellectual. But to my awareness he’s never published a political theory article. I’ll come back to this below.

AI in developmental editing

As a developmental editor, it doesn’t just peer review projects, it can suggest projects. Probably the one that folks might be the most interested in, as far as seeing what AI can do, is The Counterpublic Papers book project. Not including drafts, I’ve written 311,000 words over ten years, and have been looking for a way to generate something a bit more durable and tangible from this project. I know a volume that simply published individual issues would suck, but know there’s a way to combine these into a thicker Afrorealist rumination on life, the universe, and everything. However, I don’t have the capacity to go back over the volumes and do the work to figure out what that would look like, and while I have editor friends (some of whom read this), they lack capacity. I uploaded all the volumes to Claude and gave it a few prompts…and it developed something serviceable. Relatedly I’ve used AI to workshop my latest book project. I will say there’s variance here—I’ve used Gemini, CoPilot, and now Claude and I’ve gotten different levels of feedback. I’ve given talks on the book and have had one book workshop but the help Claude and other AIs have provided here helped me work through several blocks I had.     

AI in Writing

I needed a DJ bio last year for my set with Brett Dancer. I don’t really do self-bios well, so AI created a draft for me, that I then edited. I am turning the Detroit political imagination talk I gave earlier into a law review paper…and because I don’t generally write my talks (a byproduct of big family size—as a young parent I didn’t have time to do much more than sketch what I wanted to say), I first needed to turn that talk into a paper. An AI turned the talk transcription into the barebones of a paper. I’m in the process of turning that into something serviceable. I’ll come back to the Mounk article briefly—I think the reason he thought AI drafted a serviceable political theory paper is because he doesn’t really have political theory (and more broadly academic article) taste. And he doesn’t have taste because he neither consumes nor produces the work that would generate that taste. I suppose it’s possible to develop that taste…and maintain it without consuming or producing…but I doubt it. The more you offload consumption and production to AI, the less discriminating you become. The less taste you have.    

• Note I haven’t dealt with the political economy of AI at all. Nor have I dealt with the related issue of intellectual property rights—I participated in the Anthropic Claude class action suit, and a friend reminded me of the tragic case of the student who committed suicide after being sued because he downloaded JSTOR’s academic database, the whole thing. That’s not above my pay grade, but I don’t have the space to go into depth. What I will say, though is that I am of the belief that this is a genie in the bottle dynamic. And that the struggle, particularly given the role anti-work should play in our politics, should be around controlling AI for the public good rather than for private power and profit. 

• For more on AI in education, I’d read Hollis Robbins and Bryan Alexander. Other suggestions, please let me know. (Particularly if it looks like I forgot a newsletter.

A bit more on Prince

I re-printed my thoughts on Prince’s passing last issue. I’ll say that I’m more at peace with his passing now—I can’t imagine the pain he must’ve been in, and there’s no evidence at all to suggest he’d have stopped performing. I did some crate digging and found this, a black fantastic take.