YouTube has acknowledged that it has been applying artificial-intelligence-based processing to a subset of videos on its Shorts platform without informing creators. The company confirmed on 20 August that the test, which began several months ago, automatically unblurs, denoises and sharpens footage during upload. Rene Ritchie, YouTube’s head of editorial and creator liaison, said the project relies on “traditional machine-learning technology,” not generative AI, and is intended to improve viewing quality in a way comparable to in-camera smartphone processing. YouTube did not specify how many videos are affected, how long the trial will run or whether creators will eventually be able to disable the edits. Several high-profile creators, including musicians Rhett Shull and Rick Beato, noticed changes such as oversharpened clothing and unnaturally smooth skin and accused the service of eroding the authenticity of their work. Media experts quoted by the BBC warned that undisclosed alterations risk undermining audience trust at a time when AI-generated content is already fuelling misinformation online. The episode highlights the growing tension between platforms’ push to deploy AI-driven quality tools and content makers’ demands for control and transparency. Google, YouTube’s parent, continues to roll out other AI features across its services, even as backlash over invisible edits underscores the need for clearer consent mechanisms.