Google tests AI-powered Audio Overviews to turn search results into podcast-style summaries

New Delhi: Google is testing a new AI-powered feature called Audio Overviews, which transforms search results into conversational, podcast-style summaries. 

The experimental tool, now available through Google’s Search Labs, uses the company’s Gemini AI models to create audio recaps narrated by two AI-generated voices.

The feature, rolled out for testing on mobile devices in the US, targets nuanced queries rather than simple searches like weather updates or calendar dates. For example, searching “What is the significance of Flag Day?” or “How is Argon used in science?” may prompt a “Generate Audio Overview” button below the search bar. After a brief processing time—up to 40 seconds, according to Google—users receive a 3-to-5-minute audio clip resembling a mini-podcast, complete with two AI “hosts” discussing the topic in a conversational tone.

Audio Overviews pull information from top search results, citing sources for transparency, and allow users to adjust playback speed for convenience. The AI voices, while not perfectly human-like, are notably lifelike, featuring natural cadences, tone changes, and phrases like “Wow” or “Hold on,” though some critics note a lack of genuine rapport between the voices, likening them to podcasters reading from a script.

First introduced in Google’s NotebookLM, a research assistant tool, Audio Overviews have now expanded to Google Search, Gemini Deep Research mode, and Google Docs. The Search Labs test marks the feature’s debut in mainstream search, though it remains opt-in for now, requiring users to enable it manually. Google’s move reflects a broader push to integrate AI across its ecosystem, catering to multitaskers, auditory learners, and those with accessibility needs, such as vision impairments.

Google has not announced plans for a wider rollout but is likely gathering feedback from Search Labs to refine the feature.