Emotions and virality: why we share without verifying

In the weeks following the outbreak of the Russia-Ukraine war in February 2022, a video circulated widely across Twitter, Facebook, and Telegram purporting to show Ukrainian fighter jets in combat. The footage was real — but it was sourced from a video game. Millions of users shared it without hesitation. Not because they were careless, not because they were uninformed, but because they were afraid, angry, and invested. The relationship between emotions and virality is not incidental to information operations — it is load-bearing infrastructure. Understanding why content spreads before it is verified requires moving past content-level explanations and into the psychology and platform architecture that makes emotional amplification a reliable, exploitable mechanism. This article argues that disinformation — defined in the EU/NATO taxonomy as false or misleading content spread with intent to deceive, distinct from misinformation (spread without intent) and malinformation (true content weaponized to harm) — persists not because audiences are irrational, but because emotional arousal systematically and predictably degrades the epistemic checks that otherwise function quite well.

The neurological machinery of emotional sharing

Arousal, not valence, drives transmission

A foundational insight from Jonah Berger and Katherine Milkman’s 2012 study published in the Journal of Marketing Research — one of the most cited empirical examinations of content virality — is that it is the arousal level of an emotion, not its positivity or negativity, that predicts sharing behavior. High-arousal emotions — awe, anger, anxiety — dramatically increase the likelihood of transmission. Low-arousal emotions — sadness, contentment — do not, regardless of whether they feel good or bad. This distinction matters enormously for information operations design. State-sponsored and commercially motivated disinformation campaigns do not merely aim to make audiences feel something; they aim to make audiences feel activated.

The neurological mechanism here is reasonably well understood. Emotional arousal triggers the amygdala — the brain’s threat-detection system — which in turn compresses the timeline between perception and action. When threat-relevant content is encountered, the cognitive pathway to verification (which involves the prefrontal cortex and demands working memory, executive function, and temporal patience) is effectively deprioritized. The share button gets pressed before the source gets checked. This is not a character flaw. It is an evolved response to environmental urgency, now operating in an environment for which it was never calibrated.

The identity dimension: why we share what confirms us

Emotional engagement alone does not fully explain virality patterns. A second variable — identity congruence — substantially amplifies transmission rates. Research from the Computational Propaganda Project at the Oxford Internet Institute has documented how politically and socially identity-relevant content spreads faster within networked communities precisely because sharing functions as identity performance. When a piece of content — true or false — signals group membership, signals threat to the in-group from the out-group, or validates a pre-existing worldview, the act of sharing it becomes socially costly not to do. Verification, in this framing, is not skipped out of laziness; it is skipped because verification introduces the risk of finding out the content is false, which creates a social and psychological dilemma that many users resolve by simply not looking.

Foreign information manipulation and interference operations — the FIMI framework developed by the EU’s External Action Service (EEAS) is the most analytically precise current taxonomy — explicitly exploit this mechanism. The goal is not merely to insert false content into the information environment; it is to insert content that is emotionally and identity-resonant enough that the audience’s social network becomes the distribution infrastructure.

Why do platform architectures reward emotional activation?

Engagement optimization as an amplification mechanism

Social media platform recommendation algorithms are not neutral conduits. They are optimization systems trained primarily on engagement metrics — likes, shares, comments, watch time — because engagement correlates with advertising revenue. High-arousal emotional content produces more engagement. The algorithmic consequence is that emotionally activating content, whether true or false, receives preferential distribution. This is not a conspiracy; it is the documented output of business model incentives. The EU Code of Practice on Disinformation, updated in 2022, explicitly addresses algorithmic amplification as a structural risk factor that signatories are expected to mitigate — a recognition that the platform architecture itself is a variable in disinformation spread, not merely a passive medium.

The practical effect is a self-reinforcing loop: emotionally charged content attracts engagement, engagement signals relevance to the algorithm, the algorithm distributes it more broadly, broader distribution produces more engagement from users who encounter it in emotionally primed social contexts. By the time institutional fact-checkers or OSINT analysts — Open Source Intelligence practitioners who use publicly available information to verify claims — flag the content, it may have already completed several amplification cycles.

Cross-platform laundering and the fringe-to-mainstream pipeline

One of the more operationally significant dynamics in contemporary disinformation distribution is the fringe-to-mainstream pipeline: content originating on low-moderation platforms (Telegram channels, fringe forums, alternative social networks) migrates to mainstream platforms after achieving sufficient emotional resonance in its origin community. The Digital Forensic Research Lab (DFRLab) has documented this pattern repeatedly across influence operations tied to multiple state and non-state actors. By the time content reaches mainstream platforms, its origin is obscured, its emotional valence has been validated by early adoption, and it often carries implicit social proof from the sharing behavior of users the recipient trusts.

This cross-platform architecture is not accidental. It allows campaign operators to test emotional resonance in lower-risk environments before scaling, and it exploits the different moderation regimes across platforms as a feature rather than a vulnerability.

Myth vs. reality: are emotional sharers uniquely vulnerable?

The myth: disinformation primarily victimizes the uninformed or the partisan extreme

A persistent and analytically damaging assumption — one that occasionally surfaces even in policy documents — is that susceptibility to sharing unverified emotional content is concentrated among the poorly educated, the politically extreme, or the digitally illiterate. This framing is both empirically contested and strategically counterproductive. It leads to counter-disinformation interventions that target the wrong population and miss the actual distribution of vulnerability.

Available evidence suggests otherwise. Research published by Science in 2018 — the Vosoughi, Roy, and Aral study, arguably the most rigorous large-scale analysis of information diffusion on Twitter — found that false news spread faster, farther, and more broadly than true news, and that human behavior (not bots) was the primary driver. Critically, the study found no clear concentration of this behavior at the ideological extremes. Emotional novelty — false news was more novel and more emotionally surprising than true news — was identified as the key predictor of virality.

The reality: emotional vulnerability is broadly distributed and context-dependent

In my assessment, the more accurate framing is that emotional sharing behavior is a context-dependent failure mode that can be triggered in most users under the right conditions of arousal, time pressure, identity threat, and social priming. Security professionals, journalists, and analysts are not exempt. There is documented evidence of intelligence professionals and credentialed journalists sharing unverified emotionally resonant content during fast-moving crisis events — precisely the conditions under which arousal is highest and verification friction is most costly. The 2022 early-hours reporting on events in Ukraine produced multiple instances of credentialed journalists amplifying unverified footage, a pattern that First Draft (now part of Full Fact) catalogued as part of its crisis verification research.

This does not mean all audiences are equally susceptible across all conditions. But it does mean that counter-disinformation strategy built on a deficit model — assuming vulnerability is the property of a specific outgroup — will systematically fail to address the actual distribution of risk.

A framework for analyzing emotional amplification in information operations

Analysts and practitioners working on influence campaign attribution and impact assessment need structured approaches to evaluate how emotional amplification functions within a specific operation. The following indicators, drawn from the NATO StratCom COE analytical methodology and OSINT practice, provide a working framework:

• Arousal mapping: Identify the specific high-arousal emotional register of the content (fear, outrage, moral disgust, patriotic pride). High-arousal content in multiple registers simultaneously is a signature of professional campaign design.
• Identity resonance assessment: Determine whether the content is calibrated to specific identity communities — ethnic, religious, national, political. FIMI campaigns typically produce modular content variants targeting different identity groups simultaneously.
• Velocity and origin tracing: Use OSINT methods — reverse image search, metadata analysis, cross-platform timestamp comparison — to establish where content first appeared and how quickly it migrated. Anomalous velocity relative to the content’s apparent source is an operational indicator.
• Amplification network analysis: Map the accounts responsible for early high-volume sharing. Coordinated Inauthentic Behavior (CIB) — the term Facebook/Meta uses for organized deceptive amplification — often leaves detectable network signatures in the earliest sharing cohort.
• Emotional novelty score: Assess how surprising or emotionally unexpected the content is relative to established narrative context. High novelty content exploiting emotional expectations is more likely to bypass verification behavior.
• Friction analysis: Identify whether the content’s format and platform placement minimize verification friction — no clear source attribution, no embedded links to checkable claims, image or video format rather than text.

Conclusion: the verification problem is structural, not individual

The relationship between emotions and virality is not a problem of individual gullibility — it is a systemic feature of how information environments are currently designed and monetized. Emotional amplification is reliable enough to be engineered deliberately, and the platform architectures that reward it are not incidental to the problem but constitutive of it. Counter-disinformation approaches that focus primarily on audience education, while not without value, address only the demand side of a supply chain that has significant structural reinforcement on the production and distribution side. The more productive analytical investment is in understanding how emotional amplification functions as a force multiplier within documented information operations — and designing both platform governance and institutional response protocols accordingly.

The uncomfortable truth is that verification is not the default human behavior — sharing is. Any counter-disinformation framework that assumes otherwise is not analyzing the information environment as it exists; it is analyzing the one it wishes existed.

Sources

• Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146–1151.
• Berger, J., & Milkman, K. (2012). What makes online content viral? Journal of Marketing Research, 49(2), 192–205.
• European External Action Service. (2023). FIMI Threat Landscape Report. EEAS Strategic Communications.
• NATO StratCom COE. (2022). Cognitive Warfare: An Attack on Truth and Thought. NATO Strategic Communications Centre of Excellence.
• Bradshaw, S., & Howard, P.N. (2019). The Global Disinformation Order: 2019 Global Inventory of Organised Social Media Manipulation. Oxford Internet Institute, Computational Propaganda Project.
• European Commission. (2022). Strengthened Code of Practice on Disinformation. European Commission.