Research case
Dialogues with AI durably reduce conspiracy beliefs (Costello, Pennycook & Rand, Science 2024)
A published experiment used three rounds of personalized dialogue with GPT-4 Turbo to test changes in participants’ conspiracy beliefs, compared with unrelated-topic conversations.
What is reported
Belief fell by about 20% on average, lasted at least two months, spilled over to unrelated conspiracies, and appeared even among strong believers.
Before you use it
The result concerns this intervention and sample; it does not establish a universal persuasion effect. Research with participants requires appropriate ethics and consent procedures.
Sources and review
Source checked · Not tested by this site
Checked MIT Sloan’s detailed account of the Science study; the publisher page returned 403. This site did not replicate the intervention.
License and reuse
No license has been recorded. Check the source’s license before reusing code, models or data.
A license note is not a license audit. Open source does not guarantee free hosting, model access or commercial reuse.