TECH NEWS – The AI can read out search summaries in podcast format; Google calls this feature Audio Overview.

Google has introduced a new experimental feature called Audio Overview, which extends its AI Overview tool by delivering spoken summaries of search results in a podcast-like format. The audio is presented by two synthetic but natural-sounding AI voices. Currently, the feature is only available in the United States, supports English only, and requires manual activation through Search Labs. When enabled, the Audio Overview appears below the standard AI Overview and People also ask sections in select search results.

To hear the summary, users simply click a button. The voice content is generated live, so refreshing the page may yield an entirely different version. Each audio clip lasts between 3 to 5 minutes, featuring a back-and-forth exchange between the two voices. Although the delivery includes tone changes, conversational phrasing, and a seemingly natural style, the artificial nature is still noticeable. The voices sound like scripted podcasters with no chemistry, reading prewritten lines from a teleprompter.

Google clearly shows the sources used for each audio summary, allowing listeners to verify the information being discussed. This is important because the lifelike delivery can mislead some users into accepting everything as factual. The same issue exists with text-based AI summaries, but it’s amplified through voice. At times, one voice asks a question and then immediately answers it, exposing the artificial setup. Emotionally charged language is also used occasionally, though it often sounds stiff. Mispronunciations further betray the system’s limitations.

This feature could benefit visually impaired users or those who rely on screen readers. However, the risk of misinformation remains, and it could also threaten the livelihood of voice actors as synthetic narration becomes increasingly convincing.

Source: PCGamer, Google