Boris Mann Tech Blog

Spreading misinformation by trusting ChatGPT

September 01, 2025

Rafe Meager (they/them)'s avatar
Rafe Meager (they/them)
5mo

A case study in how Cowen's uncritical use of GPT/o3 is causing him to spread misinformation on Marginal Revolution, from @pappubahry.com pappubahry.substack.com/p/misinforma...

Misinformation on Marginal Revolution

Misinformation on Marginal Revolution

A case study in o3 hallucination


https://pappubahry.substack.com/p/misinformation-on-marginal-revolution

A long read of using ChatGPT for research, which has wrong information, and then publishing it in a blog which further cements the information even though it links only to ChatGPT output.

The author David Barry attempts to verify where the info came from and mostly debunks it. M

Subtitled, “A case study in o3 hallucination”.

Misinformation on Marginal Revolution
A case study in o3 hallucination
https://pappubahry.substack.com/p/misinformation-on-marginal-revolution

The piece ends by using the GPT-5 to debunk the previous work:

If you’ve seen the recent claim pinning this on Berthold, it appears to trace to a July 31, 2025 blog post that (unhelpfully) linked to ChatGPT as its “source,” not to any primary document. Treat that as a rumor.

David Barry concludes: “It remains bad practice to cite ChatGPT without verifying its output.”

Subscribe to Boris Mann Tech Blog
to get updates in Reader, RSS, or via Bluesky Feed
Open Source is one person
ml5.js - Friendly machine learning for the web!