Google now lets users shape their Discover feed with everyday phrases instead of rigid topic selections. The update, rolling out across the platform, replaces the traditional tap-and-choose approach with open-ended commands such as “show me more about space exploration” or “less celebrity news.”

What You Need to Know

Previously, tailoring the Discover feed required toggling specific topics or sources. The new natural language option understands intent from full sentences, making personalization more flexible. Users can adjust preferences anytime, and the system learns from ongoing inputs. The feature appears on both Android and iOS versions of the Google app.

How the Natural Language Feature Works

Users can access the new control through the Discover settings menu. A dedicated field accepts typed or spoken phrases that describe the type of content they want to see more or less of. Google’s language models interpret the request and adjust the feed accordingly.

  • Describe interests directly: Write phrases such as “more articles about climate technology” to surface related stories.
  • Reduce unwanted content: Commands like “show fewer sports updates” help narrow the feed without disabling entire topics.
  • Combine preferences: Multiple instructions can be given in a single session, allowing layered customization.

The system does not require users to learn specific keywords or categories. Google processes the meaning of the sentence and updates the recommendation algorithm in near real time.

Why This Matters

This shift hands more agency to users over a feed previously driven entirely by inferred signals. By accepting natural language, Google moves from guessing what someone wants to letting them say it directly. The change could reduce frustration with irrelevant recommendations and improve satisfaction with the Discover experience.

From a business perspective, more precise personalization may increase engagement and time spent in the app. Privacy advocates, however, will watch how Google stores and uses the free-form inputs. The company has said the language data helps train its models but has not detailed retention policies.

The update also signals a broader industry trend: major platforms are replacing nested menus and toggle switches with conversational controls. Apple, Amazon and others have introduced similar natural-language settings for their recommendation systems in recent months.

What This Means for the Personalization Landscape

Natural language customization represents a fundamental change in how users interact with algorithmic feeds. Instead of reacting to content after it appears, people now set the agenda before the feed updates. Early tests show that users issue more specific and diverse commands than when using topic lists, which may lead to more niche content surfacing.

Google has not yet expanded the feature to other personalized surfaces such as Search or YouTube. The company, however, frequently iterates on Discover features and may integrate the natural language capability more broadly later this year.