YouTube creators across gaming, technology, entertainment, and other categories reported synchronized viewership declines that creators and channel analysts trace to platform behavior changes beginning around August 13, 2025. Bellular News documented a marked desktop-to-mobile traffic reversal: desktop share fell from a 56 percent preference in July to 39.3 percent after August 13, a 16.7 percent swing toward mobile viewing. Several creators, including Skill Up and Jacob Geller, reported channel-level declines of roughly 30 percent and confirmed identical timing across multiple channels, suggesting platform-level rather than creator-driven causes.
Investigations uncovered expanded use of restricted mode filtering and new Artificial Intelligence age-estimation systems that began rolling out for U.S. users on August 13. Josh Strife Hayes found restricted mode hidden 14 of 30 recent videos on his channel despite no studio flags. Bellular Gaming showed restricted videos averaged about 250,000 views versus 421,000 for non-restricted content. Creators say restricted filtering often shows no notification in YouTube Studio, creating an information gap that prevents them from understanding or addressing distribution losses.
Creators also documented changes to engagement and monetization metrics after the August timeline. Reports showed shifts in views-to-likes ratios and views-per-like metrics, with Bellular News citing a change from 26.4 views-per-like before August 13 to 15.36 afterward. Simultaneously, reported revenue per mille increased, leading creators to hypothesize that monetizable view counts may have remained steadier while total counted views fell. Real-time analytics behavior also changed, with private real-time counts sometimes appearing lower than public metrics, reversing typical patterns and indicating possible view-count filtering or recategorization.
The period coincided with multiple platform changes, including unoriginal content detection updates in mid-July, collaboration features in early August, and confirmed machine learning video enhancement tools that some creators criticized for altering visual content. The combination of restricted mode adjustments, Artificial Intelligence age estimation, detection systems, and recommendation algorithm changes complicates diagnosis and has practical implications for sponsors, campaign planning, and creator earnings that rely on stable, transparent metrics.
