Are Trump and Xi really friends? Does Xi actually have friends? Clip from this week's interview with Ryan Hass, former White House official under Obama. Full interview link in replies.

6.) Permissible AI use will end up varying across different professors and classes. It's chaotic and confusing to students. We need shorthand descriptions in our courses. "No Tech classes, Full AI classes, etc."-- fixed labels so students know expectations before they sign up.

5.) The role of grades may need to be rethought completely. When kids are pushing for that perfect GPA, it can be detrimental to learning and classroom dynamics. I love teaching students at levels where they are purely there to learn-- MPAs, PhDs-- and grades don't really matter.

4.) AI use should be different across the academic life course. We need to protect and preserve the cognitive development of younger students, and phase in AI over time to augment the skills they've learned. We don't give kids calculators until we've taught them the multiplication tables.

3.) These new old forms of evaluation take more time-- giving oral exams to a large lecture course is a massive endeavor, which is why most professors don't do it. Ditto for grading scaffolded assignments. So if we are to do the AI thing right, we may need more professors and teaching resources.

2.) If you have high-stakes assignments that can be done with AI use, and you ban AI use without policing it, you are creating massive incentives to cheat, and that is bad for everyone. We need either assignments that embrace AI, or assignments that cannot be gamed by AI.

1.) We are starting to see real divergence across schools and programs about AI policies, ranging from the more Luddite (Berkeley Law, Chicago) to an embrace of AI tools. That may become part of schools' branding moving forward.

This is the year universities are finally tackling the AI problem in earnest. But there is no consensus across schools, or even among departments within schools, about best practices. Six things I've learned in watching this conversation evolve:

I really, really appreciate the shoutout. I am horrible at social media and growing this thing. Any suggestions or feedback on the pod are welcome!

Starting a new initiative this summer, weekly live events on Substack about the core issues facing our democracy, featuring great speakers. This Thursday, 1PM ET, an hour with the great @jasmine.bsky.social on all things work, AI, and the emerging populist backlash. Event link in replies.