YouTube’s Recommendation Algorithm Allegedly Promoted Climate Misinformation Content

YouTube’s recommendation system and its focus on views and watched time were alleged by an advocacy group to have driven people towards climate denial and misinformation videos.

Risk classification

• Primary risk domain: 3 Misinformation
• Primary risk subdomain: 3.2 Pollution of information ecosystem and loss of consensus reality

The YouTube recommendation algorithm promoted climate change misinformation at scale to millions of users, contributing to the systemic pollution of the information ecosystem and undermining consensus reality on climate science.

Additional risk subdomains

• 3.1 False or misleading information: The algorithm actively spread specific false and misleading videos about climate change to users searching for related topics.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The promotion of climate misinformation was driven by the post-deployment operations of YouTube's recommendation algorithm, which unintentionally surfaced harmful content while optimizing for user engagement.

EU AI Act risk tier

• Risk tier: 4 Minimal or No Risk

Minimal or No Risk: The YouTube recommendation algorithm is an AI system used in entertainment and content curation, which generally poses minimal direct safety risks under the EU AI Act framework.

AI system and alleged parties

• AI system: YouTube recommendation algorithm (Google)
• AI purpose: Content Recommendation; Content Curation
• Behaviour type: Autonomous
• Alleged developer: YouTube
• Alleged deployer: YouTube
• Alleged harmed parties: YouTube users, YouTube climate-skeptic users

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Severe. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Severe, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Minor
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Severe, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: Yes, the report describes the spread of climate misinformation and disinformation videos.

Directly caused: The YouTube recommendation algorithm directly promoted climate misinformation videos to millions of users through its 'Up Next' and autoplay features.

Inferred additional harm: Millions of additional users were likely exposed to toxic or misleading climate content due to the systemic reach of the recommendation algorithm.

Democracy

Reported: Yes, the report implies that the spread of climate misinformation can undermine public debate and political processes regarding climate policy.

Indirectly caused: The promotion of climate disinformation misleads the public, potentially distorting democratic debate and policy-making on environmental issues.

Inferred additional harm: Widespread algorithmic promotion of disinformation likely weakens public consensus and democratic decision-making capabilities on critical scientific and global issues.

Epistemic

Reported: Yes, the report explicitly details the spread of climate misinformation and disinformation, which misleads users about scientific consensus.

Directly caused: YouTube's algorithm recommended climate disinformation videos, with 16% of the top 100 'related videos' for 'global warming' searches containing climate disinformation, directly misleading viewers.

Inferred additional harm: The systemic promotion of false climate claims likely erodes the shared understanding of scientific reality among millions of viewers, leading to widespread epistemic harm.

People affected

• Occurrences reported: 1
• People reportedly exposed: 2000000

Potential causes

Management

• Inadequate Policy Execution: Failure to fully execute 2019 commitments to fight disinformation.
• Prioritizing Monetization: Platform profits from ads run on climate misinformation videos.

Technology

• Recommendation Algorithm Design: Algorithm promotes borderline and misinformation content to increase engagement.
• Autoplay and Up Next Features: Features automatically drive users to climate misinformation videos.

Data Inputs

• Lack of Content Classification: System fails to accurately identify and tag climate misinformation.
• Unverified Search Associations: Searching global warming links related videos containing misinformation.

Human Factors

• Advertiser Unawareness: Major brands are unaware their ads run on climate misinformation videos.
• User Vulnerability to Recommendation: Users trust and follow recommended videos without verifying accuracy.

Process and Methods

• Flawed Monetization Program: Incentivizes creators of climate misinformation by sharing ad revenue.
• Inadequate Opt-Out Tools: Places burden on advertisers to opt out of inappropriate content.
• Ineffective Borderline Reduction: Implemented methods fail to fully restrict borderline misinformation.

Regulatory Environment

• Lack of External Standards: No regulatory mandates forcing platforms to curb climate misinformation.
• Absence of Enforcement: Lack of official penalties for hosting and promoting false science.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, consistent, and specific details about the Avaaz investigation, including statistical findings (such as the 16% figure) and the mechanics of YouTube's recommendation and monetization systems. YouTube's response is also documented, providing a balanced view of the incident.
• Ambiguities identified: The exact inner workings and optimization metrics of YouTube's recommendation algorithm are proprietary and not fully detailed in the reports.
• Alternative interpretations: The incident could be viewed primarily as a policy/moderation enforcement failure rather than an algorithmic failure, as the algorithm was performing its designed task of maximizing engagement.

The YouTube recommendation algorithm autonomously promoted climate change misinformation to millions of users to maximize watch time. While this represents a significant scale of information pollution and minor societal risk, it does not pose an immediate threat to critical infrastructure, sovereignty, or national security operations.

• Overall national security impact: Minor
• Scope: Multiple nations
• Primary target: No clear primary
• Other affected: Multiple nations
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. Represents an ongoing systemic issue regarding algorithmic curation rather than an active crisis requiring immediate intervention.
• Autonomy: Full autonomy. The recommendation algorithm autonomously curates and serves video content to users in real-time without human oversight for individual decisions.
• Novelty: Established threat. Algorithmic amplification of controversial or misleading content to drive user engagement is a well-documented and established industry issue.

Impact by dimension

• Physical security: Negligible. No direct threats to physical systems, critical infrastructure, or human safety were reported in connection with the recommendation algorithm.
• Information security: Minor. The algorithm systematically promoted climate misinformation to millions of users, representing a minor scale of information manipulation, though lacking evidence of state-sponsored coordination.
• Sovereignty: Negligible. No direct threats to state authority, territorial control, or core government operations, though a US Senator requested policy changes.
• Economic security: Negligible. No threat to strategic industries or financial systems, though some major brands had ads displayed on misinformation videos without consent.
• Societal stability: Minor. The promotion of misinformation erodes public consensus on critical global issues like climate change, posing a minor long-term risk to societal decision-making.