AI Chatbots Pose Risk of 'Knowledge Collapse' by Narrowing Information Diversity, Studies Warn

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

Multiple studies led by Stanford University and the University of Copenhagen reveal that AI chatbots like ChatGPT, Gemini, and Claude provide users with significantly less diverse information than traditional web searches. Researchers warn this could lead to 'knowledge collapse,' reducing access to varied perspectives and potentially harming societal knowledge diversity.[AI generated]

Why's our monitor labelling this an incident or hazard?

The article explicitly involves AI systems (large language models powering chatbots) and discusses a plausible future harm (knowledge collapse) that could arise from their use. No actual harm has been reported as having occurred; the study warns about potential risks if AI-generated content becomes dominant and reduces information diversity. This fits the definition of an AI Hazard, as the event plausibly could lead to harm (harm to communities through reduced access to diverse knowledge) but does not describe a realized incident. The article also provides recommendations to mitigate this risk, but these are part of the study's findings rather than a response to an incident.[AI generated]
AI principles
Transparency & explainabilityDemocracy & human autonomy

Industries
Media, social platforms, and marketing

Affected stakeholders
General publicConsumers

Harm types
Public interestHuman or fundamental rights

Business function:
Other

AI system task:
Interaction support/chatbotsContent generation


Articles about this incident or hazard