Master's Thesis

Fact-Checking News with Internet Memes

Sérgio Miguel Gonçalves Pinto2025

Key information

Authors:

Sérgio Miguel Gonçalves Pinto (Sérgio Pinto)

Supervisors:

Daniel Jorge Viegas Gonçalves (Daniel Jorge Viegas Gonçalves); Helena Sofia Andrade Nunes Pereira Pinto (Helena Sofia Andrade Nunes Pereira Pinto)

Published in

05/26/2025

Abstract

Technological advances such as the internet and social networks disrupted the news ecosystem by reducing constraints on the dissemination of news, compromising journalistic norms of objectivity and ethics. Traditional fact-checking approaches face scalability and slowness issues, reaching only a fraction of users, and often arriving after peak dissemination. Recent literature suggests emotionally evoking content as a potential avenue for effective changes in people’s beliefs. Given memes’ low cognitive effort, humor load, and virality potential, we investigate their effectiveness as a medium for fact-checking. We developed MemeFact, a Retrieval Augmented Generation system that draws contextual information from the Internet Memes Knowledge Graph to automatically generate fact-checking memes. A user study with 110 participants revealed that memes generated by our MemeFact system, using Claude-3.5-Sonnet, surpass human-created ones in coherence, clarity, and persuasiveness metrics, while falling short significantly in hilarity. Our LLM-as-a-judge approach found weak alignment between AI evaluations and human judgment, with AI models systematically providing significantly more positive ratings. Crucially, a study with 313 participants demonstrated that meme-based explanations reduced backfire effects (39.60% compared to 58.18% for textual explanations) while achieving higher rates of false belief correction (49.50% vs 31.82%). Providing contextual information alongside memes proved critical in enhancing credibility. On X, our meme-based posts for correcting misinformation resulted in 7.56% of users deleting their false content. These findings suggest that memes offer a promising pathway for propagating corrections through social networks by counterbalancing the cognitive affective discomfort of belief challenge while effectively correcting false beliefs.

Publication details

Authors in the community:

Supervisors of this institution:

Fields of Science and Technology (FOS)

electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering

Publication language (ISO code)

eng - English

Rights type:

Embargoed access

Date available:

03/29/2026

Institution name

Instituto Superior Técnico