Does Misinformation and Deepfake Content Create Measurable “Trust Stress”?

Direct answer: Yes, and the mechanism is different from ordinary misinformation. Researchers analyzing 1,231 real tweets discussing deepfakes during the Russia-Ukraine conflict found genuine, authentic videos were mislabeled as fake 60 times, far more often than actual deepfakes were correctly identified, which happened only 11 times. The core finding across this research area is that deepfakes aren’t necessarily more persuasive than other misinformation, their real damage is epistemic: they erode the general capacity to trust visual evidence at all, genuine footage included.

Why Researchers Distinguish “Trust Damage” From Simple Persuasion

This distinction matters for understanding what deepfakes actually do differently from older forms of misinformation. Research reviewing how political deepfakes affect audiences found they don’t carry a meaningfully higher inherent persuasive power than other misleading content, people aren’t necessarily more convinced by a fake video than by a fake headline. What deepfakes do more distinctively is produce epistemic destabilization, a broader erosion of confidence in the entire category of visual evidence, precisely because visual proof has traditionally been treated as uniquely trustworthy, and that specific trust is what gets undermined once convincing forgeries are known to be possible.

Why Real Content Gets Mislabeled as Fake Far More Than Fake Gets Correctly Caught

This is the study’s most concrete and counterintuitive finding. Among the 1,231 tweets analyzed, drawn from an initial pool of 4,869 collected during 2022, genuine, authentic conflict footage was accused of being a deepfake 60 separate times, while actual deepfakes were correctly identified only 11 times. That five-to-one imbalance shows the practical, lived effect of deepfake awareness isn’t sharper detection of real fakes, it’s a broader, less discriminating suspicion cast over real evidence far more often than it catches genuine manipulation.

Why Awareness of Deepfakes Backfires Into Broader Distrust

The same analysis documented a related pattern researchers call the “liar’s dividend,” the phenomenon where simply knowing convincing fakes exist makes it easier to dismiss real, inconvenient evidence as fabricated. The dataset included 155 separate instances of conspiracy theories claiming specific world leaders were themselves deepfakes or had been secretly replaced, a direct, real-world illustration of how awareness that deepfakes exist can be weaponized to discredit authentic information rather than only protecting against fake information.

Why the Behavioral Response Is Withdrawal, Not Sharper Skepticism

A further, important pattern from the broader research base: frequent exposure to manipulated political content doesn’t reliably make people better, more careful evaluators of evidence, it tends to reduce their overall confidence in the information environment and motivate them to disengage entirely, avoiding political news and discussion rather than developing more refined judgment. Researchers describe this specific outcome as epistemic fatigue, a state of exhausted disengagement rather than the improved discernment a more optimistic account might predict.

Why This Compounds Regardless of How Media-Literate Someone Is

A separate survey spanning eight countries found prior exposure to deepfakes increased belief in unrelated misinformation, and notably, this pattern held for social media news consumers regardless of their general cognitive ability. That’s a meaningful finding against a common assumption, that better critical thinking skills or media literacy alone would fully protect someone from this effect, the erosion documented here appears to operate somewhat independently of how sophisticated a given consumer’s reasoning otherwise is.

What This Means for Understanding Misinformation’s Psychological Toll

The practical takeaway is that deepfakes and related misinformation create a genuine, documented form of stress distinct from simply being fooled by a false claim, a broader erosion of trust that spills over onto real, authentic evidence far more often than it sharpens detection of actual fakes. Given that the documented behavioral response tends toward withdrawal and fatigue rather than improved skepticism, and that this effect isn’t fully offset by general cognitive ability, treating this as purely an information-literacy problem likely understates what the research actually shows about its psychological and social cost.

Sources: The 1,231-tweet thematic analysis of deepfake discourse during the Russia-Ukraine conflict, including the 60-versus-11 mislabeling finding and the 155-instance “liar’s dividend” pattern, sourced directly from PMC, “Do deepfake videos undermine our epistemic trust? A thematic analysis of tweets that discuss deepfakes in the Russian invasion of Ukraine.” The epistemic-destabilization framing, the epistemic-fatigue behavioral pattern, and the eight-country cognitive-ability finding cross-checked against ScienceDirect, “Exploring how political deepfakes trigger cognitive processes and downstream psychological and behavioral outcomes: A systematic review.” Verified 2026-08-08.