Science Communication and Truth Perception
Truth Evaluation Mechanisms
Truth evaluation relies on two primary processing pathways: analytical evaluation and feelings-based processing.
Analytical evaluation assesses five core criteria (Schwarz, 2015; Schwarz et al., 2016):
Credibility
Coherence
Consistency
Consensus
Evidence
"Truthiness" describes truth originating from gut feeling rather than facts or books, a concept introduced by Stephen Colbert on Comedy Central's The Colbert Report in October 2005.

Cognitive Fluency and the Illusory Truth Effect
Cognitive processing ease (fluency) serves as a key metacognitive cue informing truth judgments, causing easily processed information to feel true.
Illusory Truth Effect: Repeating information increases processing fluency, causing repeated claims to be judged as truer than novel claims.
Holds true even when individuals possess prior demonstrated or estimated knowledge to the contrary (Fazio et al., 2015; Fazio, 2021).
Pre-exposure warnings decrease overall truth ratings but do not eliminate the relative truth boost for repeated claims over new claims (Jalbert et al., 2020; Naderevic & Abfalg, 2016).
Persists regardless of individual Need for Cognition (NFC) levels or the specific nature of the claim (Newman et al., 2021; Jiang et al., 2022).

Forms of Fluency Impacting Perception:
Perceptual Fluency: High visual clarity (Reber & Schwarz, 1999) and higher audio quality significantly elevate ratings of talk quality, speaker intellect, speaker liking, and research importance (Newman & Schwarz, 2018).
Linguistic Fluency: Phonological ease (easily pronounceable names or terms) enhances perceived trustworthiness (Silva et al., 2017; Laham et al., 2012; Newman et al., 2014); complex, high-level vocabulary reduces evaluations of competence (Oppenheimer, 2006).
Conceptual Fluency: Presenting non-probative photos alongside claims increases belief by enhancing imagery and conceptual processing ease (Derksen et al., 2022; Newman et al., 2012, 2021).

Online Spaces and Synthetic Content Vulnerabilities
AI Hyperrealism: Synthetic AI-generated human faces are frequently evaluated as more realistic and trustworthy than actual human photos due to representing "averages of averages" (prototypes) that cluster densely around the center of face-space (Miller et al., 2023; Nightingale & Farid, 2022).

Manipulative Media: Generative AI tools enable widespread deceptive media, such as deepfake videos (e.g., manipulated videos of Ukrainian President Volodymyr Zelenskyy or fake doctor supplement promotions) and synthetic social media influencers.
Information Ecosystems: Digital spaces filled with self-aligned content, in-groups, and trusted sources induce "cognitive cruising," heightened susceptibility, and lowered vigilance regarding claim validity (Kozyreva et al., 2020; Newman et al., 2022; Oyserman & Dawson, 2020).
Misinformation Interventions
Pre-emptive Skills: Media literacy education and inoculation/prebunking strategies build resilience against deceptive tactics (Guess et al., 2020; Lewandowsky & van der Linden, 2021; Roozenbeek et al., 2022). Transparent AI labeling provides necessary contextual awareness.

Environmental Friction & Nudges: Accuracy nudges prompt individuals to deliberate on statement correctness, significantly reducing sharing intentions for false content (Butler et al., 2024; Fazio, 2020). Over 80% of U.S. adults across political affiliations agree it is very important to share only accurate content online.
Corrections & Debunking: Debunking and AI-leveraged corrections reduce belief in misinformation, with periodic reminders further strengthening intervention long-term efficacy (Ecker et al., 2022; Spearing et al., 2025; Costello et al., 2025).
Third Person Effect: Individuals tend to perceive misinformation as having a greater negative impact on others ("everyone else" > "friends and family") than on themselves ("me") (Yang & Tian, 2021).