Overview
- YouTube’s Rene Ritchie confirmed an experiment that applies machine learning to selected Shorts during processing to reduce noise and increase sharpness, and he said generative AI is not involved.
- Creators report visible artifacts such as oversharpened details, odd hair and skin rendering, and subtle distortions, and they say they were neither notified nor allowed to decline the edits.
- Following backlash, Ritchie said YouTube is working on an option that will let creators disable the automatic post-processing on their Shorts.
- Reporting from outlets including the BBC and Heise has characterized the post-processing as generative AI, a description YouTube disputes by framing the work as traditional machine learning.
- The episode has revived questions about YouTube’s 2023 disclosure policy for manipulated or synthetic media and whether platform-applied changes require labeling or an opt-out.