Google Tests AI-Powered "Tailor Your Feed" to Personalize Discover Content

Pasukan Editorial BigGo
Google Tests AI-Powered "Tailor Your Feed" to Personalize Discover Content

Google is pushing the boundaries of content personalization with a new experiment that puts users in the driver's seat. Through its Search Labs program, the company is testing a feature called "Tailor your feed," which allows users to directly instruct their Google Discover feed using plain language. This move represents a significant shift from algorithmic curation to user-directed AI, aiming to solve the common problem of irrelevant content cluttering personalized news streams.

A New Level of Control for Discover Users

The core of the "Tailor your feed" experiment is a simple text input box within the Google app. Here, users can type specific requests about the content they wish to see or avoid. This goes beyond the existing controls, which allow users to hide topics or sources. For example, a user could instruct Discover to "only show me vegan dessert recipes" or "keep me updated on space exploration but avoid commercial launch news." Google's AI interprets these prompts to refine the feed's content sourcing, potentially filtering by publisher, content format, or even abstract qualities like the desired "vibe" of the articles.

Feature Name: Tailor your feed Platform: Google app on Android (U.S. only) Access Method: Search Labs experiment Core Function: User provides plain-language text prompts to customize Google Discover feed content. Example User Prompts: "stop showing me negative news," "keep me updated on country music," "only show recipes for gluten-free diets." Customization Levers: Can target or exclude specific topics, publishers, creators, content formats, and abstract qualities ("vibes"). Interface: Activates a chat-like thread with an AI to confirm and refine changes.

Integration and Access Through Search Labs

This feature is not yet a standard part of the Google app. It is currently a limited experiment available within Search Labs, Google's platform for testing early-stage search and feed features. Users in the United States who have enrolled in Search Labs on Android may see a new lab flask icon in their Google app. Activating the "Tailor your feed" experiment from this menu enables the functionality. The process involves typing a request, reviewing the AI's interpretation of the desired changes, and confirming them, allowing for iterative refinement of the feed.

The Context of Google's Evolving Discover Feed

This test is part of Google's ongoing effort to make its Discover feed more engaging and sticky. Discover, which appears on the homepage of the Google app and on many Android devices, aims to surface relevant articles and videos without a user needing to search. Earlier in 2025, Google experimented with "Daily Listen," a feature that converted a user's feed into a synthesized audio podcast. Both experiments point toward a future where Google's AI doesn't just find content but actively repackages and personalizes it, potentially keeping users within Google's ecosystem rather than clicking out to source websites.

Related Google Discover Experiment (2025):

  • Feature Name: Daily Listen
  • Function: AI-generated audio summaries ("podcasts") of content from a user's Discover feed and Search history.
  • Implication: Part of a trend toward AI repackaging content within Google's ecosystem, potentially reducing direct clicks to external websites.

Implications for Content Discovery and Privacy

The "Tailor your feed" feature highlights the growing role of generative AI in content curation. By understanding natural language requests, the system promises a more intuitive and powerful way to manage information intake. However, it also raises questions about the "filter bubble" effect, where algorithms narrowly tailor content to confirmed interests, and about data privacy, as users provide explicit details about their preferences. The success of this experiment will likely depend on how accurately Google's AI can fulfill diverse and nuanced user requests while maintaining a sense of serendipity and breadth in the content it surfaces.