Kuznetsova, V. (2026). All talk, no rules: Content moderation practices and discursive strategies in platform governance of climate misinformation. Platforms & Society, 3. https://doi.org/10.1177/29768624261499168
Summary
This article investigates how four major social media platforms — Meta, YouTube, TikTok, and X/Twitter — govern climate misinformation both in their formal rules and in their public self-representation. Through qualitative content analysis of 95 platform documents, Kuznetsova shows that climate misinformation is rarely codified as a distinct, high-priority moderation category, even as platforms make vocal public commitments about addressing it. Platforms favor “soft” interventions (labeling, demotion, demonetization) over removal, and deploy a recurring repertoire of discursive strategies that cast themselves as benevolent, neutral infrastructures while redistributing the actual labor and responsibility of governance onto users, creators, and external experts. The paper frames this gap as “governance-washing” and argues that platform self-governance has structural limits requiring external, enforceable oversight.
Key Contributions
- Systematic cross-platform comparison of how climate misinformation is both formalized (in policy) and operationalized (in public communications).
- Extends platform governance research from policy monitoring into the discursive dimension, naming three specific legitimation strategies in the under-studied climate case.
- Empirically documents the misalignment between platform rhetoric and codified/enforced rules.
- Positions climate misinformation as a distinctive governance puzzle relative to COVID-19 and election misinformation, using TikTok to show stricter climate rules are operationally feasible.
- Connects findings to regulatory debate, arguing the DSA systemic risk framework could but has not covered climate misinformation.
Methods
Qualitative content analysis of 95 public-facing platform documents (policies, advertising rules, community guidelines, and public communications), collected in early 2025 via structured Google Advanced searches and direct navigation. RQ1 used thematic coding classifying documents as “explicit,” “implicit,” or “not mentioned,” drawing on Schaffner et al.’s elements of comprehensive content moderation and Goanta’s documentation framework. RQ2 used inductive, constant-comparison qualitative coding (Braun and Clarke) to surface recurring discursive strategies of justification and responsibility framing across platforms.
Findings
- All four platforms discussed climate misinformation in public communications, but formal policy integration was uneven or absent; only TikTok has explicit rules in both content and advertising guidelines.
- TikTok is the only platform that removes user-posted climate misinformation, using a tiered system; Meta, YouTube, and X/Twitter rely on demotion, labels, demonetization, or minimal treatment.
- Definitions anchor in “scientific consensus” but remain abstract and require external knowledge; concrete examples and exceptions are inconsistently provided.
- Three cross-platform discursive strategies: (1) “increasing the good, decreasing the bad” (justifying soft moderation via counter-speech and engagement metrics); (2) “empowering others” (externalizing epistemic authority to experts and NGOs); (3) “authorizing content creators” (shifting advocacy responsibility onto creators).
- Scientific authority is selectively operationalized — invoked to define misinformation but withheld as grounds for removal.
- Neither scientific consensus nor political contestation plausibly explains the absence of climate-specific rules, since COVID-19 and election misinformation received dedicated rules including removal.
- Governance is volatile: Meta discontinued US fact-checking and deactivated its Climate Science Center in 2025 while leaving outdated documentation online.
Connections
This paper’s focus on platform governance mechanisms, policy enforcement, and the gap between stated and codified rules connects it to work auditing platform rule-making and content moderation infrastructure such as Rieder2025-ju, Rieder2026-pp, and Hurcombe2025-cs. Its emphasis on climate-specific misinformation and intervention design situates it alongside the message-intervention literature in that register, including Spampatti2026-kx, van-der-Linden2026-jt, and Lewandowsky2026-ob. The regulatory framing around external oversight and the DSA links it to governance and data-access scholarship such as de-Vreese2026-zx and Bechmann2026-dr.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify