Paper Club #012: "Pitfalls of Evidence-Based AI Policy" by Casper, Krueger, and Hadfield-Menell (2025)
For the folks who are already familiar with us, here's this week's paper:
Title: Pitfalls of Evidence-Based AI Policy
Author(s): Stephen Casper, David Krueger, Dylan Hadfield-Menell
Date published: Feb 2025
Link to paper: https://arxiv.org/abs/2502.09618
Here's the link for the suggested papers moving forward. Learn more about us below.
What is the Paperclip Minimizer Club?
Join us for an hour-long discussion about a paper relevant to AI safety. Our intention is to collectively unpack key ideas, identify points of confusion, and deepen our understanding of the field.
How does it work?
A volunteer facilitator will guide us through a discussions of the paper. Mainly, what we want to understand is:
Big Idea: What is the big idea for this paper? Or the main point the authors are trying to get across?
Results: What are the key findings and achievements?
Relevance: Why does this research matter? What problem does it address? i.e. How does this progress the field?
Related work: How does it improve on previous work?
Open Questions: What open-ended questions do you have after reading this piece?
How can I participate?
Message Lenz to join. Just send a note introducing who you are and why you want to join.
Pre-read the paper. Attendees are encouraged to read the paper in advance and come prepared with questions or insights. If you need tips on how to read a paper quickly, watch this lecture by Andrew Ng.
RSVP to this event. Make sure to let us know that you're coming. We're currently limiting the discussion to just 10 people to see how it goes.
Suggest papers. Feel free to add more papers through this link!