Confronting Misinformation Produced with Generative AI: A Meta-Analysis of Experimental Scientific Evidence

The report finds that textual GenAI misinformation is increasingly perceived as credible over time, while visual GenAI misinformation, such as deepfakes, shows a decline in perceived credibility. Preventive corrective information and content labeling are identified as effective countermeasures, though their impact varies by context.

• Textual generative AI misinformation shows a statistically significant upward trend in perceived realism and credibility between 2019 and 2024, with a modest increase in effect size over time (β = 0.08).
• Visual generative AI misinformation, including deepfakes, has experienced a statistically significant decline in perceived credibility since approximately 2021 (β = -0.22).
• Preventive corrective information—such as educational forewarnings about AI hallucinations or deepfake techniques—is associated with a consistent reduction in the perceived credibility of GenAI misinformation (g = -0.30 for post-2020 studies).
• Content labeling (e.g., tags like "AI-Generated") provides a small but statistically significant reduction in the perceived credibility of GenAI misinformation (g = -0.24), though its effectiveness is highly context-dependent.
• The meta-analysis is based on 60 effect estimates extracted from 24 peer-reviewed publications, focusing exclusively on randomized controlled trials (RCTs).
• The report identifies a significant geographic gap in research, as the analyzed evidence is heavily concentrated on participants from the USA and the European Union.

APA

Herasimenka, A., Valenzuela, S., Boulianne, S., Esser, F., Given, L. M., Lewandowsky, S., Navarro-López, E. M., & Howard, P. (2026). Confronting Misinformation Produced with Generative AI: A Meta-Analysis of Experimental Scientific Evidence. International Panel on the Information Environment. https://www.ipie.info/research/sr2026-2-confronting-generative-ai-misinformation-meta-analysis

Chicago

Herasimenka, Aliaksandr, Sebastián Valenzuela, Shelley Boulianne, Frank Esser, Lisa M. Given, Stephan Lewandowsky, Eva M. Navarro-López, and Philip Howard. Confronting Misinformation Produced with Generative AI: A Meta-Analysis of Experimental Scientific Evidence. International Panel on the Information Environment, 2026. https://www.ipie.info/research/sr2026-2-confronting-generative-ai-misinformation-meta-analysis.

Wikipedia

{{cite report |last1=Herasimenka |first1=Aliaksandr |last2=Valenzuela |first2=Sebastián |last3=Boulianne |first3=Shelley |last4=Esser |first4=Frank |last5=Given |first5=Lisa M. |last6=Lewandowsky |first6=Stephan |last7=Navarro-López |first7=Eva M. |last8=Howard |first8=Philip |title=Confronting Misinformation Produced with Generative AI: A Meta-Analysis of Experimental Scientific Evidence |publisher=International Panel on the Information Environment |date=2026 |url=https://www.ipie.info/research/sr2026-2-confronting-generative-ai-misinformation-meta-analysis |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{herasimenka2026confronting, author = {Herasimenka, Aliaksandr and Valenzuela, Sebastián and Boulianne, Shelley and Esser, Frank and Given, Lisa M. and Lewandowsky, Stephan and Navarro-López, Eva M. and Howard, Philip}, title = {{Confronting Misinformation Produced with Generative AI: A Meta-Analysis of Experimental Scientific Evidence}}, institution = {International Panel on the Information Environment}, year = {2026}, url = {https://www.ipie.info/research/sr2026-2-confronting-generative-ai-misinformation-meta-analysis}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }