User intent

The assessment of whether users spread false information intentionally (with knowledge or reckless disregard for falsity) or unintentionally (misled and unaware). Intent is foundational to platform policy and intervention strategies.

Scope and motivation

Intentional spreading includes: - Malicious disinformation campaigns (coordinated propaganda, state actors, troll farms) - Commercial fraud and scams - Harassment and manipulation for personal gain

Unintentional spreading includes: - Misled by false information before sharing - Confirmation bias and selective exposure leading to unwitting amplification - Repeated exposure increasing perceived accuracy (validity effect)

Policy implications: - Intentional spreaders → suspension, account removal, coordinated inauthentic behavior detection - Unintentional spreaders → gentle interventions: nudging (encouraging accuracy checks), fact-presenting recommendation algorithms, connection to fact-checkers

Key papers and articles

• Zhou et al. (2022) — "This is Fake! Shared it by Mistake": Assessing the Intent of Fake News Spreaders: First systematic quantitative assessment of user intent via influence graphs capturing internal (user's past posts) and external (other users' posts) psychological and social influences. Introduces influence graph with edge weights based on article/content similarity and temporal decay. Validates on MM-COVID and ReCOVery datasets with manual and algorithmic annotations. Shows assessed intent significantly improves fake news detection (AUC ~0.93 with hand-crafted features, 0.83–0.91 with graph neural networks) and can distinguish between bots/trolls (intentional) and misled users (unintentional).
• Wang et al. (2025) — InSIDE: Intent-Semantic Joint Learning for Fake News Detection: Combines intent signals (underlying purpose/motivation of articles) with semantic signals via heterogeneous graphs. Shows consistent improvements on four datasets; outperforms prior intent-aware methods.

Connections

• Fake news detection — intent assessment improves detection accuracy
• Bot detection — key signal for distinguishing bot/troll accounts from misled humans
• Social media and misinformation — platform-level interventions should be intent-aware
• Misinformation — unintentional spreading driven by psychological factors (confirmation bias, availability cascade, social identity)
• Propagation-based fake news detection — influence patterns and network position reveal spreading intent

Open questions

• Can affect/emotion be reliably inferred from social media posts to improve intent assessment?
• Do cross-platform account histories improve intent modeling?
• How effective are nudging and fact-presenting interventions at changing unintentional spreader behavior?
• Can intent assessment be real-time on platforms without compromising user privacy?