SSRN Electronic Journal · 2026

Persuadable Regardless: Conspiracy-Relevant Traits May Predict Whether AI-Induced Belief Change Lasts, Not Whether It Occurs

Nikola Erceg et al., with Jonas R. Kunst

In brief

A preprint examining how AI chatbots change people's conspiracy beliefs and whether individual traits linked to conspiratorial thinking shape that process. The title suggests a distinction between whether AI-induced belief change happens at all and whether it endures: conspiracy-relevant traits may predict the durability of updating rather than its initial occurrence, implying that most people are persuadable in the moment regardless of such traits.

Read the full article ↗ DOI: 10.2139/ssrn.7661495