Google rolls out conversational AI customization for the Discover feed
Google has begun rolling out a real, conversational way to customize the Discover feed — users can now type a plain-language request for what they want to see more or less of, and the feed adjusts and remembers that preference going forward, rather than relying purely on implicit engagement signals.
Why this is a genuinely different customization model
Most real feed-personalization systems work implicitly — inferring preferences from what a user clicks, scrolls past, or lingers on. A direct, conversational customization option is a genuinely more explicit mechanism, letting a user state a preference directly rather than hoping the algorithm eventually infers it correctly from behavior alone.
The real, broader pattern this fits
This is part of a genuine, ongoing trend of large consumer platforms adding conversational, natural-language controls on top of existing recommendation systems — a real, practical application of language models beyond chat interfaces, applied directly to an existing, high-traffic product surface.
For publishers and content creators relying on Discover traffic, a real shift toward explicit, user-stated preferences (rather than only implicit engagement signals) could genuinely change what content actually surfaces — worth watching how this rollout affects real traffic patterns over the following weeks.
Source: 9to5google.com