Particle.news

Download on the App Store

YouTube Confirms Quiet Machine-Learning Edits on Some Shorts, Prompting Creator Backlash

The company says the tests use non‑generative techniques to reduce blur and noise, but creators are pressing for clear disclosure and an opt‑out.

Photo: Olly Curtis/Future via Getty Images (Getty Images)
A person holds a phone in front of the YouTube logo.
An image about YouTube is quietly using AI to change some videos without creator consent
The YouTube play logo is being displayed on a smartphone, with YouTube in the background, in this photo illustration in Brussels, Belgium, on July 24, 2024. (Photo Illustration by Jonathan Raa/NurPhoto via Getty Images)

Overview

  • Rene Ritchie of Team YouTube said select Shorts are being processed with traditional machine learning to unblur, denoise, and boost clarity, not generative AI or upscaling.
  • Creators including Rhett Shull and Rick Beato posted side‑by‑side comparisons showing smoother skin, sharper fabric details, and occasional ear or edge warping on YouTube versus other platforms.
  • Reports trace user complaints to June, with BBC, Ars Technica, Gizmodo, and others documenting the alterations as media scrutiny intensified in late August.
  • Affected creators argue unannounced edits undermine authenticity and audience trust, and they are calling for the feature to be optional with clear labeling.
  • YouTube likens the processing to modern smartphone enhancements and says it is gathering feedback, but it has not announced creator controls or an opt‑out.