Google Experiments with Audio Answers on Search with Gemini AI | Generative ai in finance and banking | What is generative ai | Microsoft generative ai tools | Turtles AI

Google has introduced “Audio Overviews” in Google Search via Labs: short audio summaries, generated by Gemini models, with an integrated player and links to sources, for a hands-free experience. The initiative raises questions about the impact on publisher traffic.

Key points:

• New Audio Overviews feature available in Labs.
• Audio player with controls, speed, and links to sources.
• Uses Gemini models, podcast-style virtual hosts.
• Potential reduction in traffic to publishers.

Google has started a test in the United States, integrating an experimental feature called “Audio Overviews” into Google Search for mobile and desktop. Based on Gemini, this new feature allows users who have activated Labs to automatically generate audio synths – lasting about four minutes – for certain queries, such as “how do noise-canceling headphones work?”. The activation is done via a “Generate Audio Overview” button under the “People Also Ask” section, with generation times of up to 40 seconds.

The audio player, embedded directly in the results page, offers play/pause functions, volume control, listening speed (from 0.25× to 2×), and links to the original sources. There is also a thumbs-up or down vote for each overview, which Google uses to receive feedback and improve the experience. The audio features two virtual host voices, conversing in a podcast-like style.

The new feature is an evolution of the experience introduced in NotebookLM and the Gemini app, where the feature offers podcast-style audio generation from documents, legal notes, or shared courses. On Search, the goal is to expand accessibility and meet the needs of auditory learners or those who prefer a hands-free approach.

The political and industrial context of Google’s move comes at a time when several analyses, including a report by the Wall Street Journal, highlight how AI summaries—textual and now also audio—are drastically reducing traffic to publishers’ sites. According to the WSJ, publications such as HuffPost, Washington Post, Business Insider and New York Times have seen a decline in search referrals of between 50% and 55%, with drastic cuts in staff. Click-through rates have dropped by 40–60% on many articles when an AI Overview appears. At the same time, Google defends the positive impact on the “discoverability” of content, even if users’ attention is increasingly shifting towards immediate and concise answers.

According to Matt Southern of Search Engine Journal, even if Google maintains transparency by citing sources within the audio player, it remains to be seen whether the links will be enough to counteract the decline in visits to the original sites. In an evolving landscape, publishers are exploring new avenues, such as direct licensing models, agreements with AI companies or strategies based on subscriptions and partnerships to recover lost margins.

From a technological point of view, the implementation confirms Google’s commitment to making search increasingly multimodal: voice is joined by text, images, interactive modes such as AI Mode, Search Live and advanced voice discovery, all presented during I/O 2025. However, challenges remain, such as the management of AI "hallucinations": already with text summaries, cases were highlighted in which the system generated incorrect or even dangerous information (such as using glue in sauce), or invented non-existent proverbs.

Google expands its search offering with synthetic audio responses, to meet new learning styles and hands-free usage situations. The road map includes feedback collection via Labs and integration only if the experiment is perceived as useful. Only time will tell whether the function will be able to compensate for the risks of reduced traffic to original content and how the balance between practicality and attribution of sources will evolve.

A final reflection: the arrival of Audio Overviews marks a crucial step in the way we consume content, offering advantages of accessibility and immediacy, but it forces publishers to rethink distribution and monetization models in an ecosystem increasingly driven by AI.