Dataconomy
  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Whitepapers
  • AI toolsNEW
  • Newsletter
  • + More
    • Glossary
    • Conversations
    • Events
    • About
      • Who we are
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
Subscribe
No Result
View All Result
  • AI
  • Tech
  • Cybersecurity
  • Finance
  • DeFi & Blockchain
  • Startups
  • Gaming
Dataconomy
  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Whitepapers
  • AI toolsNEW
  • Newsletter
  • + More
    • Glossary
    • Conversations
    • Events
    • About
      • Who we are
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
Subscribe
No Result
View All Result
Dataconomy
No Result
View All Result

Google Discover tests AI headlines that rewrite news with errors

Users have noticed factual inaccuracies, such as a headline claiming "BG3 players exploit children."

byKerem Gülen
December 5, 2025
in News, Tech
Home News
Share on FacebookShare on TwitterShare on LinkedInShare on WhatsAppShare on e-mail

Google has implemented a new artificial intelligence (AI) experiment that rewrites news headlines for some users in Google Discover, leading to concerns about accuracy and potential misinformation. The firm confirmed the new headlines are part of a “small UI experiment for a subset of Discover users.”

The feature impacts content on Google Discover, which users access by swiping right on Pixel or Galaxy home screens, scrolling in the Google app on iPhone, or opening a new Chrome browser window with Google as the homepage. AI-generated article previews now sometimes replace headlines and descriptions written by authors and editors.

Evidence of the initiative first appeared when The Verge staff writer Sean Hollister experienced the altered headlines. While some AI headlines appeared vague, others presented factual inaccuracies, according to reports. For instance, an AI headline stated, “Steam Machine price revealed,” though the original Ars Technica article indicated Valve’s Steam Machine would not be priced like a typical console. Another AI headline implied “Qi2 slows older Pixels,” while the original article merely stated that older Pixel phones cannot use the full extent of Qi2 fast-charging capabilities.

Stay Ahead of the Curve!

Don't miss out on the latest insights, trends, and analysis in the world of data, technology, and startups. Subscribe to our newsletter and get exclusive content delivered straight to your inbox.

Concerns have also emerged regarding the AI’s capacity to generate misleading or potentially libelous headlines. In one instance, a PCGamer article with the original headline “Child labor is unbeatable” was rewritten by Google’s AI as “BG3 players exploit children,” presenting a significantly different and potentially harmful implication.

Google previously introduced similar AI previews in October, which included short AI summaries that users could expand. The current experiment, however, changes the placement of headlines, positioning the AI-generated text where original headlines typically appear.

Users can identify AI-generated headlines by several characteristics:

  • Length: AI-generated headlines are frequently four words or fewer.
  • Capitalization: These headlines often capitalize only the first word, diverging from standard editorial style guides.
  • Attribution: Users can tap “See more” under the Discover preview to find a tag indicating “Generated with AI, which can make mistakes.” Articles with original headlines do not feature this button.

Google has confirmed no opt-out mechanism for this AI experiment is available to users at this time.


Featured image credit

Tags: Featuredgoogle discover

Related Posts

Anthropic partners with Teach For All to train 100,000 global educators

Anthropic partners with Teach For All to train 100,000 global educators

January 20, 2026
Signal co-founder launches privacy-focused AI service Confer

Signal co-founder launches privacy-focused AI service Confer

January 20, 2026
Adobe launches AI-powered Object Mask for Premiere Pro

Adobe launches AI-powered Object Mask for Premiere Pro

January 20, 2026
Google Workspace adds password-protected Office file editing

Google Workspace adds password-protected Office file editing

January 20, 2026
Claim: NVIDIA green-lit pirated book downloads for AI training

Claim: NVIDIA green-lit pirated book downloads for AI training

January 20, 2026
Tesla restarts Dojo3 supercomputer project as AI5 chip stabilizes

Tesla restarts Dojo3 supercomputer project as AI5 chip stabilizes

January 20, 2026

LATEST NEWS

Anthropic partners with Teach For All to train 100,000 global educators

Signal co-founder launches privacy-focused AI service Confer

Adobe launches AI-powered Object Mask for Premiere Pro

Google Workspace adds password-protected Office file editing

Claim: NVIDIA green-lit pirated book downloads for AI training

Tesla restarts Dojo3 supercomputer project as AI5 chip stabilizes

Dataconomy

COPYRIGHT © DATACONOMY MEDIA GMBH, ALL RIGHTS RESERVED.

  • About
  • Imprint
  • Contact
  • Legal & Privacy

Follow Us

  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Whitepapers
  • AI tools
  • Newsletter
  • + More
    • Glossary
    • Conversations
    • Events
    • About
      • Who we are
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
No Result
View All Result
Subscribe

This website uses cookies. By continuing to use this website you are giving consent to cookies being used. Visit our Privacy Policy.