factual

LLMs cannot study misinformation due to guardrails

LLM-based agent models cannot be used to study misinformation dynamics because providers like OpenAI prevent the models from producing misinformation, forcing such study to rely on real social media data instead.

factualpending

Speaker

Petter Törnberg

Evidence Quote

“in part because the LLMs are under... OpenAI doesn't let them produce misinformation, so you can't really use them to study that.”

Source

330 | Petter Törnberg on the Dynamics of (Mis)Information— Sean Carroll's Mindscape
Created: 6/13/2026, 12:22:44 AM

Your notes

Loading notes...