Snapchat has moved to limit the spread of fully AI-generated videos in its Spotlight feed, the discovery section that surfaces trending clips to users who do not already follow the creators. As of the end of July, videos produced entirely by artificial intelligence will no longer appear in recommendations. The platform frames the change as an effort to keep the space focused on work that begins with real people rather than automated systems that can churn out repetitive material at scale.
The rule does not ban AI-generated clips from the app altogether. Users may still post them in other places. Creators can also keep using Snapchat’s own AI features to edit or enhance footage they shot themselves, and those mixed videos remain eligible for Spotlight as long as transparency indicators are present. The distinction rests on whether a human supplied the underlying recording or whether the entire video was synthesized. In an environment where synthetic clips have grown harder to tell from filmed ones, that line is both practical and increasingly difficult to police. Snapchat itself has noted that no detection system is perfect.
Recommendation algorithms matter because they determine whether a piece of content reaches beyond an existing audience. For automated operations that flood platforms with near-identical videos, losing that distribution channel removes much of the incentive. Spotlight has drawn more contributors in the past year, with the number of unique people posting to it rising by more than 120 percent. Keeping the feed oriented toward human-originated work is one way to protect that growth from being diluted by low-effort synthetic volume.
Other platforms have begun drawing similar boundaries, each in its own way. LinkedIn has expanded tools for members to flag posts that feel like automated filler and is adjusting its recommendation systems to reduce the visibility of generic AI text. Substack offers readers an optional detector that estimates how much of a longer post or comment was written by a person versus assisted by machines. YouTube continues to withhold monetization from repetitive or mass-produced inauthentic material, while TikTok has tested controls that let individuals dial the amount of AI-generated content they see. These moves reflect a shared recognition that an endless supply of formulaic synthetic media can erode the sense that anything on a feed is worth genuine attention.
Meta, by contrast, has described a future in which its models generate a nearly infinite stream of personalized images and videos that sit alongside posts from friends and established creators. The contrast is instructive. One set of companies is trying to preserve pockets of human-first discovery; another is preparing to treat synthetic content as a primary recommendation source. Snapchat has chosen the narrower path for Spotlight, betting that users still value work that starts with a person even as the tools for full generation grow cheaper and more capable. Whether the detection methods prove reliable enough to enforce that preference remains an open question, but the preference itself is now explicit.
