Sign in

Google's big action on 'odd and misleading search results' in AI Overviews: What you need to know

Google will introduce limits on AI Overviews after it delivers misleading and inaccurate search results

Published on: May 31, 2024, 14:47:05 IST
Share
Share via
  • facebook
  • twitter
  • linkedin
  • whatsapp
Copy link
  • copy link

Google is introducing limits on AI Overviews, its AI search experience, after it delivered several odd and inaccurate results to consumers, the company said on May 30, Moneycontrol reported.

The Google logo is seen on the Google house (REUTERS)
The Google logo is seen on the Google house (REUTERS)

A CNET article stated that when asked how to keep cheese on pizza, it suggested adding an eighth of a cup of nontoxic glue. The AI took this from a 11-year-old Reddit comment originally meant as a joke. Also, it recommended eating "at least one small rock per day," from a 2021 satirical piece of The Onion.

This is a form of AI hallucination, which occurs when a generative AI model serves up false or misleading information and presents it as fact. Hallucinations result from flawed training data, algorithmic errors or misinterpretations of context, according to the CNET article.

Also Read | Google Search algorithm secrets revealed by whistleblower

Google's search head Liz Reid said the company has now built detection mechanisms for "nonsensical queries" and has limited the inclusion of satire and humor content, in a blogpost on May 30, Moneycontrol reported.

Google spokesperson Meghann Farnsworth said the mistakes came from “generally very uncommon queries, and aren’t representative of most people’s experiences.” The company has taken action against violations of its policies, she said, and are using these “isolated examples” to continue to refine the product, The Verge reported.

Also Read | Google techie says getting rejected for promotion was ‘blessing in disguise': 'I knew I needed to show…'

This is the second major Google AI launch this year that has come under scrutiny for delivering inaccurate results as the image generation feature of Gemini, Google’s competitor to OpenAI’s ChatGPT quickly came under fire for outputting historically inaccurate images like Black Vikings, racially diverse Nazi soldiers and a female pope, forcing Google to issue an apology and pause the feature, according to a Forbes article.

These developments also come at a time when Google is looking to monetise its AI offerings, by testing search and shopping ads in AI overviews, according to the Moneycontrol report.

Also Read | Google leases new office in Bengaluru for 4 crore monthly rent: Report

  • Abhyjith K. Ashokan
    ABOUT THE AUTHOR
    Abhyjith K. Ashokan

    Deeply passionate about writing, Abhyjith works as a business journalist covering corporates, markets, the economy, and policy - forces that in many ways, shape the world and pave the path for intriguing storytelling. For him, breaking news is a high that only gets matched by the adventures of the open road; both of which he deems essential to what matters at the end of the day. The story.Read More

Stay updated with the latest Business News, stock market updates, petrol and diesel prices, gold and silver rates, income tax updates and major developments from India and across the world.