If you thought something looked slightly off on your YouTube videos these past few weeks, you’re not going crazy. In a move that should surprise absolutely nobody, Google has been putting AI where you didn’t ask for it — secretly editing your uploads because, well, it can. All for the sake of improving upload quality, of course.
A not-so-secret experiment
At least, that’s what Rene Ritchie, YouTube’s head of editorial, had to say on the subject in a post on X. The experiment hit “select” YouTube Shorts earlier this year, using “traditional machine learning” to unblur, denoise, and improve clarity in videos. A valiant goal, but still one that didn’t need to be kept from the public.
“No GenAI, no upscaling…” Ritchie said. “YouTube is always working on ways to provide the best video quality and experience possible, and will continue to take creator and viewer feedback into consideration as we iterate and improve on these features.”
Read More: YouTube is giving free users one of its best Premium features
Putting aside the fact that machine learning is a form of AI (though technically not generative AI), YouTube’s little experiment hasn’t gone quite according to plan. Users have reported several errors with some Shorts these last few months, with many pointing out the too-smooth edges, weird artefacts, and a generally ‘AI look’.
This is what caused YouTuber Rehtt Shull to begin looking deeper into the issue after noticing similar errors on his Shorts, eventually leading him to make this video on the subject. It became impossible for Google to pretend it wasn’t altering videos, though it still remains unclear whether this will be the new standard going forward, or if creators will be given the option to turn off AI-enhancements before publishing.
