Google Discover is getting an AI chatbot-tuned feed
Google is rolling out a new AI-powered customisation tool for its Discover feed, allowing users to describe in natural language what content they want to see. The feature, arriving in the Google app "in the coming days," uses a chatbot-style interface to let users state their preferences, which the system will then remember and apply automatically to future visits, giving people more direct control over the recommendations that appear based on their Search and app activity.
Users will access the tool via the three-dot menu on their Discover feed, where a chatbot window lets them describe preferences, review a summary of what the AI will prioritise, and refine the request before hitting "Refresh your feed." The move follows similar AI-driven feed customisation features already introduced by YouTube, Instagram, Bluesky and X. Google also announced related updates, including personalised topic selection for Google News audio briefings on Android and an enhanced "Preferred Sources" feature letting publishers add an interactive button so readers can favourite outlets directly from a webpage.
- Google Discover will let users describe preferences via an AI chatbot
- Feature remembers preferences and refreshes the feed accordingly
- Google also updated News audio briefings and Preferred Sources features
New here? Start with this
Google Discover is a personalised feed in the Google app that suggests articles and other content. Its recommendations can be shaped by a person’s Search and app activity, as well as by topics and sources they follow.
Google makes the app and the systems that choose what appears in Discover. Publishers produce the articles and other content that may be recommended, while readers can use feed settings to influence what they see. Similar tools for guiding recommendations have appeared on platforms including YouTube, Instagram, Bluesky and X.
These feeds affect how people encounter news and other information online. Giving readers more ways to state their preferences may change the balance between their choices and automated recommendations.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Proponents argue this feature genuinely improves user autonomy by making algorithmic curation more transparent and intentional. Rather than systems passively inferring preferences from behaviour, users can now explicitly state what they want to see, creating a clearer and more honest relationship between choice and outcome. Natural language interfaces make such control accessible to everyone, not just those comfortable with technical settings, whilst respecting that different users have different information needs. This shift from implicit to explicit preference-setting represents a meaningful step towards reducing algorithmic opacity.
The case against
Critics contend the feature risks deepening filter bubbles by making it frictionless for users to isolate themselves in narrow information silos, potentially worsening polarisation and reducing exposure to viewpoints that challenge assumptions. They argue the notion of 'user control' obscures that Google's underlying business model still depends on extensive data collection about preferences and behaviour, and that the algorithm—not users—ultimately determines what appears in the feed. The real structural problem, in this view, is that algorithmic feeds prioritise engagement over informational quality, and marginal improvements to user interface preferences do nothing to address that fundamental misalignment.