The News, Explained
Netflix announced AI Discover on November 3 as a new experimental space for finding titles in its TV app. It is rolling out to a small group of US members who use Netflix in English on supported TVs, with access expanding over the coming weeks. The change concerns how viewers explore recommendations, making eligibility and feedback controls useful starting points. Source
The first experience begins with themes. Options reflect a mood, company or available time, such as a light watch, a family movie or something short, and are previewed with clips. Selecting a theme brings up matching titles. Viewers can narrow the selection through more specific themes or choose Something Different for another set. Netflix says AI helps create the themes and that the experience explains its results. Source
The US Help Center adds eligibility details. AI Discover is optional, available to selected members and excluded from Kids profiles. It requires a TV or TV streaming device that meets the technical requirements. Included members see a Discover shortcut in the TV’s top menu and receive an email. If the shortcut is missing, the member may be outside the trial or the device may not meet the requirements. This help article gives no model-by-model list or numerical device requirements. Source
The scope of recommendation feedback is explicit. Choosing Don’t recommend under a title’s Tune Recommendations control removes that title from recommendations within AI Discover. Recommendations elsewhere on Netflix stay unchanged. For example, excluding a title in this experimental space applies when exploring within AI Discover. The US help article describes an optional experience alongside the regular Netflix experience. Source
The experiments may evolve. Netflix says usage and feedback will inform which experiences it keeps or improves, and some features could eventually enter the main Netflix experience. The help article also warns that AI-powered experiences can make mistakes. Source Source
OYOPICK’s Take
We are particularly interested in viewers being able to reject an unwanted title and understand where that choice applies. Themes provide a starting point for narrowing the selection; title-level feedback offers a way to adjust subsequent exploration. If both work well together, they could help when a conventional genre label does not capture what someone wants to watch. Whether viewers actually find suitable titles more easily still needs evidence from the trial.
The AI Discover-only scope matters for understanding why different screens may behave differently. As the trial grows, we hope viewers can clearly see that scope and easily reverse a choice. Clear explanations of recommendations and their controls could give viewers more room to use a new discovery system according to their own judgment.