Google is testing a Dive deeper button in Discover that opens a short topic overview with links to related coverage

Google has initiated a new experimental feature within its Discover feed, introducing a "Dive deeper" button designed to provide users with a more comprehensive understanding of trending topics. Announced by Robby Stein, Google’s Vice President of Product for Search, via a social media update on X, the feature represents the latest iteration in the company’s ongoing strategy to integrate generative artificial intelligence into the user experience of its mobile feed. Initially launching as a test focused exclusively on video content, the feature aims to bridge the gap between passive content consumption and active information gathering by surfacing summaries and related reporting directly within the Google app.

The Mechanics of the Dive Deeper Experience

The "Dive deeper" functionality functions as a gateway to an AI-generated overview. When a user encounters a video card in their Discover feed—such as a news clip or a weather report—they are presented with a concise, AI-crafted summary of the subject matter. Below this summary, the new button invites users to expand their exploration.

According to the visual documentation provided by Stein, the interface unfolds in a multi-layered fashion. Upon tapping the button, the user is transitioned to a dedicated screen headed by the current topic. This interface aggregates several key components: a more detailed AI-generated summary, a curated selection of "source cards" linking to original reporting, and community reactions. By grouping these elements, Google is attempting to mitigate the echo-chamber effect often associated with algorithmic feeds, ensuring that users are exposed to a broader spectrum of perspectives and primary sources.

Crucially, the feature includes a disclaimer, explicitly noting that the AI-generated summaries are subject to potential inaccuracies. This aligns with Google’s broader risk-mitigation strategy regarding Large Language Model (LLM) outputs, which are prone to "hallucinations" or factual errors. By clearly labeling these segments, Google seeks to maintain user trust while leveraging the efficiency of automated content synthesis.

Chronology of AI Integration in Google Discover

This test follows a significant trajectory of AI deployment within Google’s ecosystem. The evolution of Discover from a simple interest-based feed to an AI-assisted information hub has been incremental but steady:

• October 2025: Google formally rolled out AI-powered previews for trending topics in the Discover feed. These previews were designed to expand upon a topic, offering users a quick synthesis of information coupled with outbound links to reputable web content. The initial launch targeted the United States, South Korea, and India, serving as a primary testing ground for international reception.
• August 31, 2026: Google expanded the accessibility of its Generative AI performance report within Search Console. This tool was developed to provide web publishers with visibility into how their content performs when surfaced through AI-driven features in Discover.
• September 2026: Google provided further clarification on the scope of its Search Console reporting, confirming that impressions originating from generative AI features would be tracked, though it maintained a distinction between general experimental features and those operating within the Search Labs environment.
• Current Phase: The introduction of the "Dive deeper" button represents a shift toward interactive, rather than just informative, AI features. Unlike the previous iteration, which merely expanded a summary, this feature creates a navigational hub for a specific topic.

The Role of Generative AI in the Search Ecosystem

The integration of generative AI into Discover is not merely a user-interface enhancement; it is a fundamental shift in how Google processes and distributes information. By generating summaries, Google is effectively acting as a curator, a role that traditionally belonged to news aggregators and journalists.

This creates a complex dynamic for the web ecosystem. On one hand, Google argues that these features increase the discoverability of original journalism by providing direct links to source material. By placing a "source card" at the top of the "Dive deeper" overview, the company aims to drive traffic back to the original publisher’s domain. On the other hand, there is an ongoing industry debate regarding whether these summaries satisfy the user’s intent to the point where they no longer feel the need to click through to the original website.

For publishers, the "Dive deeper" feature represents a new variable in their search performance. Since the feature is currently in an experimental stage, Google has not provided granular data regarding how traffic from these specific "Dive deeper" cards will be categorized in Search Console. If the current trend holds, impressions will likely be bundled into the broader "generative AI features" category, making it difficult for publishers to isolate the specific impact of this new interaction model on their site traffic.

Analytical Implications for Content Discovery

From a product design perspective, the "Dive deeper" button serves as a "call to action" (CTA) that encourages engagement. In the attention economy, the primary metric for success is the amount of time a user spends within the Google app. By providing a "Dive deeper" overview, Google keeps the user engaged with the interface for a longer duration while simultaneously providing a path toward deeper reading.

However, the efficacy of this feature rests on the quality of the AI synthesis. If the summaries provided in the "Dive deeper" screen are consistently accurate and provide high-value links, users will likely adopt the feature as a primary way to consume news. Conversely, if the summaries appear superficial or fail to provide a diverse range of perspectives, the button may see low adoption rates.

Stein has emphasized that Google is currently in a period of iterative design. The company plans to test multiple layouts and interaction patterns in the coming weeks. This suggests that the current manifestation—a video-centric, summary-heavy screen—is subject to change based on engagement metrics like click-through rates (CTR) and user retention.

Official Stance and Future Roadmap

Google has remained characteristically opaque regarding the specific rollout schedule or the geographic scope of this test. There is no mention of when the feature might move beyond video content or whether it will be integrated into other search verticals.

The lack of a timeline is consistent with Google’s "test and learn" methodology. By starting with a narrow segment (videos) and a limited user base, the company can measure the impact on both user experience and publisher metrics without disrupting the core Search or Discover products.

Furthermore, the separation of this experiment from Search Labs is noteworthy. While many of Google’s more radical AI features are incubated in Search Labs, "Dive deeper" appears to be an evolution of the existing Discover product, suggesting that Google views AI-enhanced content discovery as a core component of the future user experience rather than a peripheral experimental feature.

Conclusion: A New Standard for Information Consumption

The "Dive deeper" initiative underscores Google’s commitment to consolidating its position as the primary gateway for digital information. As AI continues to change how users interact with content, the line between an aggregator and a content creator continues to blur.

For the average user, this feature offers a streamlined way to move from a single video to a curated, multi-source understanding of a topic. For the industry, it presents a challenge to adapt to a world where the search engine is no longer just a list of blue links, but a generative layer that sits between the reader and the publisher. As Google continues to iterate on this feature, the tech community will be watching closely to see if the "Dive deeper" button successfully enhances the value of the open web or if it further centralizes power within the Google ecosystem. For now, the experiment serves as a clear indicator that the future of search is conversational, contextual, and increasingly synthesized.