YouTube secretly tested AI video enhancement without notifying creators

A wave of discontent is sweeping through the YouTube creator community as many have noticed something peculiar about their recent video uploads. After a detailed investigation by a well-known music channel, it has been revealed that Google has been conducting secret tests on a new feature that employs AI to enhance video quality. While the company asserts that this initiative aims to deliver superior video experiences, the lack of communication to creators about the experiment has raised significant concerns. The testing phase, which began earlier this year in YouTube Shorts, has attracted scrutiny due to reports of unusual artifacts, edge distortion, and an overly smooth appearance in videos—hallmarks of AI manipulation. Creators have likened this effect to the unwanted oversharpening seen when zooming in on smartphone photos. Rene Ritchie, YouTube's head of editorial, clarified the technology's nature in a post on X (formerly known as Twitter), stating that the feature utilizes "traditional machine learning" rather than generative AI. This approach aims to minimize blur and noise while enhancing image sharpness, yet it still involves AI in the video modification process. YouTuber Rhett Shull, who began probing the changes after discussions with peers, is convinced that these AI enhancements, which he refers to as "upscaling," have been implemented without adequate notification to creators. Despite Ritchie's insistence that this is not technically upscaling technology, the controversy over the lack of transparency continues to grow.