Ex-AI researcher explains why the fake-news flood prediction never came true

neil_chilson · x · 2026-09-26

A researcher who pivoted from AI research to studying why past AI predictions failed pushes back on the idea that she left the field.

When ChatGPT launched, many thoughtful people considered the release reckless: models could write fake news in seconds, personalized propaganda would flood the internet, and society would lose the ability to tell truth from falsehood. That never happened.

Her argument: the forecast wrongly assumed people believe what they see. In reality, people trust those they know and beliefs their side already holds. Roughly 1% of users see most fake content online, mostly those actively seeking it, and the share of conspiratorially minded people has stayed stable throughout history.

Takeaway: predicting how the world will change requires understanding how it actually works now—not how headlines, politicians, or your own field frame it.

Original post →

More from AGI Musings

AGI Musings channel →