SAN FRANCISCO: Google has begun rolling out a new artificial intelligence feature in Google Docs that generates short, podcast-style audio summaries of documents, the company said.
The feature, called “Audio Summaries”, is powered by Google’s Gemini AI model and is being introduced to paid Workspace users as part of what the company described as an extended rollout starting on Feb. 12.
The update adds a new option under Tools > Audio > Listen to Document Summary.
When selected, Gemini analyses a document’s content – including material across multiple tabs – and produces a spoken recap that typically lasts less than three minutes.
AI-generated synopsis
Unlike conventional screen readers that read text word-for-word, the new tool synthesises information, according to Google.
It extracts key ideas and presents them in a natural, conversational tone.
The approach is similar to the “Audio Overview” feature in Google’s NotebookLM product, which provides spoken summaries of source material.
Google said the tool generates a script based on the core concepts of a document rather than simply converting text to speech.
Designed for multitasking
The feature includes several customisation options. Users can choose between different voice personas, including a neutral “Narrator”, an energetic “Persuader” and a supportive “Coach”.
A built-in media player allows listeners to pause, rewind and adjust playback speed between 0.5x and 2x.
The company said the feature is aimed at professionals reviewing reports and students working through long notes, offering a way to consume information while commuting or multitasking.
Limited to paid users
Audio Summaries are currently restricted to paid subscribers and are not available for free personal Google accounts.
Eligible plans include Google AI Pro and Ultra tiers, Business Standard and Business Plus, Enterprise Standard and Enterprise Plus, as well as accounts with Google AI for Education or Business add-ons.
Google described the rollout as “extended”, meaning it may take longer than the usual 15 days for the feature to reach all eligible users.
