Google's AI-Generated Headlines in Discover Feed Spread Misinformation
Google tested replacing news headlines in its Discover feed with AI-generated ones that were often inaccurate, misleading, and stripped of context from the original reporting. The experiment resulted in headlines like 'BG3 players exploit children' and 'Steam Machine price revealed' that did not match the actual article content.
What happened
Google began experimentally replacing original news headlines with AI-generated ones in its Discover feed as of December 2, 2025. The AI-generated headlines often stripped tone, accuracy, and context from original stories, producing misleading or inaccurate results.
Reported impact
Examples of the misleading AI headlines included "BG3 players exploit children," "Steam Machine price revealed" (when Valve had not revealed pricing), "AMD GPU tops Nvidia" (misrepresenting AMD announcements), and "Microsoft developers using AI" (omitting context that made the original headline meaningful). Users reading Google Discover noticed the strange and inaccurate headlines appearing on their feeds.
Index analysis
The headlines were reduced to four words or less, often losing critical context and nuance from the original reporting. While some AI-generated headlines were acceptable, most were misleading or contained factually inaccurate claims not supported by the underlying articles.
Governance controls
Google disclosed the effort as "a small UI experiment for a subset of Discover users" but provided little transparency to readers that AI was rewriting the headlines they were seeing.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.