How to Retrain Your YouTube Feed and Remove 'AI Slop'
Manage activity signals and watch history to reclaim your homepage from irrelevant recommendations.
YouTube users often find their home feeds cluttered with irrelevant content or 'AI slop' as their interests shift or shared accounts skew data. To fix a broken feed, users must manually nudge the platform's recommendation engine to ensure it remains an efficient discovery tool.
The recommendation algorithm relies heavily on activity signals to determine what a user wants to see. To begin the process, users must first ensure that watch history is enabled in their Google Account settings. If this feature is disabled, YouTube cannot generate personalized suggestions, which can result in a completely blank home feed.
Training the Algorithm
To retrain the system, users should provide explicit positive and negative signals. Positive signals—which tell the algorithm to show more of a specific type of content—include watching videos to completion, liking content, and leaving comments. These actions signal a high level of engagement and interest to the AI.
Conversely, users can reduce irrelevant content by utilizing negative signals. The most effective method is using the three-dot menu to select 'Not interested' or 'Don't recommend channel.' Disliking videos and unsubscribing from outdated channels also help prune the feed of unwanted topics.
Managing History and Access
One of the most effective ways to refine recommendations is through the 'Manage YouTube History' tool. Rather than performing a full wipe of all data—which removes all context for the AI—users should selectively delete specific videos. This allows for the removal of one-off views or accidental clicks that do not reflect core interests.
To prevent future skewing, especially on shared devices, using Incognito mode or setting up separate accounts for different users is essential. These methods prevent 'incorrect watching'—such as a child using a parent's account—from polluting the primary user's recommendation data.
Why Feed Control Matters
Because the volume of content uploaded to YouTube is so massive, the vast majority of users rely on the algorithm for discovery rather than manual searching. When the system fails, the platform transforms from a discovery engine into a source of frustration. By actively managing these signals, users can reclaim their digital environment and ensure the content they see aligns with their current preferences.
While these manual adjustments provide immediate relief, the long-term effectiveness of specific signals remains a point of interest. Users should continue to monitor how the platform responds to selective history deletion versus active feedback tools like the 'Not interested' button to determine the fastest path to a curated feed.