YouTube’s Recommendation Algorithm Allegedly Promoted Climate Misinformation Content

February 1, 2019

A recent investigation by an advocacy group has revealed that YouTube's recommendation algorithm may be inadvertently promoting climate misinformation content to viewers. The focus on views and watched time in the algorithm can sometimes lead to climate denial and misinformation videos being pushed to users, despite their expressed interest in more accurate information. This can create a cycle of misinformation that is difficult to break. For example, if someone watches a video promoting climate denial, YouTube may then recommend similar content based on their viewing history, potentially further solidifying their misconceptions about the issue. To address this concern, it is essential to examine the impact of the algorithm on the dissemination of misinformation and explore ways to improve its recommendations to promote more accurate information.

Matched TAIM controls

Suggested mapping from embedding similarity (not a formal assessment). Browse all TAIM controls

Alleged deployer
YouTube
Alleged developer
YouTube
Alleged harmed parties
YouTube users, YouTube climate-skeptic users

AI governance case studies

For forensic AI governance failure analysis (TAIMScore™ case studies), browse Human Signal’s Failure Files™.

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/305

Data source

Incident data is from the AI Incident Database (AIID).

When citing the database as a whole, please use:

McGregor, S. (2021) Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database. In Proceedings of the Thirty-Third Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-21). Virtual Conference.

Pre-print on arXiv · Database snapshots & citation guide

We use weekly snapshots of the AIID for stable reference. For the official suggested citation of a specific incident, use the “Cite this incident” link on each incident page.