🚨New Paper w/ @joelleforestier.bsky.social at JEP: General!🚨 We conducted 2 mega-experiments totaling over 28,000 participants and 50 conditions. We wanted to find the most effective interventions to reduce implicit weight prejudice across 5 implicit measures. 🧵 osf.io/preprints/ps...

"Implicit" bias + AI research often studies biased outputs. As us psychologists know though, behavior's not the same as process. A model trained on a racist site would show bias, but wouldn't be "implicit"! To study & find biased processing, @lee-messi.bsky.social & I used new reasoning models.

In a new paper w/ @calvinklai.bsky.social, I find that OpenAI’s latest reasoning model (o3-mini) exhibits implicit bias-like patterns. What’s exciting about reasoning models is the ability to unpack bias in how models *process* information, rather than just seeing bias in *outputs*. (1/10):