Media release
From:
Countering AI-Generated Misinformation With Pre-Emptive Source Discreditation and Debunking
Despite concerns over AI-generated misinformation, little research has examined its impact on reasoning and the effectiveness of countermeasures. Across two experiments, a misleading AI-generated article influenced reasoning regardless of its alleged source (human or AI). A source-discreditation that highlighted issues with AI reliability reduced general trust in AI-generated information but did not reduce the article’s impact on reasoning. A simple disclaimer also had no impact. A content-focused correction effectively reduced misinformation influence, but only a combination of content- and source-focused interventions eliminated it entirely. Findings demonstrate that AI-generated misinformation can be persuasive, potentially requiring multiple countermeasures to negate its effects.