Particle.news

Download on the App Store

Google Tests AI-Generated Podcast-Style Audio Summaries in Search

Powered by Gemini models, it delivers short spoken recaps through two lifelike voices citing top webpages.

Google Search has introduced Audio Overviews to help users listen to select search results aloud using AI-powered podcast-style summaries.
The Google logo and lettering can be seen on the façade of the company's Munich headquarters
The Google Gemini logo
blank

Overview

  • Google rolled out the Audio Overviews feature in its Search Labs as an opt-in experiment for U.S. users, generating spoken summaries for nuanced or complex queries.
  • The tool is powered by Gemini models and uses two AI-generated voices that play off each other in a podcast-style format.
  • Audio summaries range from three to five minutes long and include citations to the web pages used to compile the information.
  • Availability is currently limited to English content and requires users to toggle it on within Search Labs.
  • Advocates note its usefulness for users with vision impairments, and critics caution that its lifelike voices could facilitate misinformation and undercut voice acting roles.