Good morning, folks! It’s another week full of AI developments. As I remind my actual students—the demanding pace of AI progress means everyone is constantly learning about the promises and pitfalls of this emerging technology. If you meet an “AI expert,” remind them that there’s no such thing because we’re all novices in this time of AI madness.
A few assignments this week. Be sure to share them with a colleague. Homework is easier with a study group, right?
First: Learn about Mutual Assured AI Malfunction (MAIM)—a deterrence strategy akin to Mutual Assured Destruction that aims to reduce the odds of superintelligent AI inciting a horrific conflict.
Easy assignment - listen to my podcast interview with Dan Hendrycks, who co-authored the paper with Eric Schmidt (former Google CEO) and Alexandr Wang (Scale AI CEO).
Challenge assignment - read their whole paper, which includes an overview of MAIM as well as other strategies.
Second: Vice President J.D. Vance further refined the administration’s AI policy position in a speech at the American Dynamism Summit. Vance’s remarks expanded upon key themes he outined back in Feburary at the Paris AI Action Summit. Given that Congress has been slow to pick up AI legislation and the White House has been dealing with some other issues, Vance’s speeches offer the best insights as to what we can expect from administration’s impending AI Action Plan (forthcoming in July).
Easy assignment: check out this analysis of Vance’s speech as well as my assessment of the “America First, America Only” approach to AI that I inferred from his remarks.
Challenge assignment: read the whole speech here.
Third: More than 700 bills have popped up in state legislatures. The substance of these bills varies wildly from a “Right to Compute” in Montana to measures that would drastically restrict the ability of labs to train their models on copyrighted data. Making sense of this regulatory fervor is easier said than done.
Easy assignment: hear from Dean Ball and Sunny Ghandi on the ongoing regulatory developments in states.
Challenge assignment: Read up on your own state’s proposals and, if you really want some extra credit, consider sending your legislator some thoughts on whether its a step in the right direction!
Extra Credit:
Attend this fascinating conversation hosted by The Gov Lab on misinformation and AI (tonight at 5pm ET)
Read through OpenAI’s submission in response to the request for information related to the AI Action Plan. This important filing gives us a sense of how at least one major lab thinks about the proper role of the federal government in regulating AI.
Finally, for all those who celebrate, happy March Madness! Unfortunately, the Longhorns have already found their way out of the Dance, but the Oregon Ducks will take on Liberty in the First Round.