Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Geoff Hinton and his disciples have never wavered in their beliefs and fears of …
ytc_UgySKGvox…
G
New professions and services will emerge based on human emotions, which AI can't…
ytc_Ugx2CQ956…
G
We appreciate your feedback and understand your concerns. If you're interested i…
ytr_UgxeTvJt9…
G
Replacing 5 people with AI 👎
Replacing 5 people with 1 person who uses AI 👍…
ytc_Ugymsido4…
G
ChatGPT, like any other technology, is not without its potential negative effect…
ytc_UgzhF3ZzY…
G
Are you familiar with the first chat bot ELIZA?
It is absolutely common for us …
rdc_nk43uhg
G
umm, LLM usually stands for large language model, not language learning model, y…
ytc_UgzO7J7Ur…
G
I do not feel sorry for the boy or his family tbh, why would you ask an AI about…
ytc_UgxJF9NJ0…
Comment
This is what Grok says. I'm also aware this video is several months old so it relates to the times mentioned by Grok.
---++
Yes, YouTube did automatically apply AI-based enhancements to some uploaded Shorts videos without the creator's knowledge, request, or consent during a 2025 experiment.
Key Details
What happened: Starting around June 2025, YouTube ran an experiment on select YouTube Shorts where it automatically applied "traditional machine learning" (a form of AI) during video processing to unblur, denoise, sharpen, and improve clarity. This altered the visual appearance of videos post-upload, often resulting in a "smeary," "oil painting-like," over-smoothed, or plastic-looking effect on skin, backgrounds, and details.
Automatic and without permission:
Creators did not opt-in, were not notified in advance, and only discovered it when comparing their Shorts to versions on other platforms (e.g., Instagram) or noticing viewer complaints. Many creators, like Rhett Shull and Rick Beato, publicly called it out, emphasizing it eroded trust by changing their original work without consent.
YouTube's response:
In August 2025, after backlash, YouTube's creator liaison confirmed the experiment and described it as similar to smartphone auto-enhancements. They insisted it was not generative AI (no new content created) but traditional ML for quality improvement. Following criticism, they promised (and later introduced) an opt-out option for creators.
Scope and status:
It was limited to a subset of Shorts (not all videos or all Shorts), aimed at improving mobile viewing quality. There's no evidence this became a permanent, universal feature by late 2025; later updates focused on creator-controlled AI tools or TV super-resolution. The experiment appears to have been scaled back or made opt-out due to feedback.
In summary, yes—YouTube did auto-enhance some Shorts without uploader involvement in mid-2025, sparking significant controversy over transparency and creative control. Most other AI features on Shorts (like text-to-video or effects) are opt-in tools for creators to use themselves.
youtube
2025-12-31T12:2…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | company |
| Reasoning | unclear |
| Policy | unclear |
| Emotion | indifference |
| Coded at | 2026-04-26T19:39:26.816318 |
Raw LLM Response
[
{"id":"ytc_Ugwoc2OukyW4pCiuXHN4AaABAg","responsibility":"company","reasoning":"unclear","policy":"unclear","emotion":"indifference"},
{"id":"ytc_UgxqeQZA_XyVxAqoOYF4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"industry_self","emotion":"approval"},
{"id":"ytc_Ugz0ymEljEvf0Udk5554AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugy3cgStvtkKAplxjb54AaABAg","responsibility":"company","reasoning":"unclear","policy":"unclear","emotion":"fear"},
{"id":"ytc_Ugz5mHeuoyn5kQ1_gBR4AaABAg","responsibility":"company","reasoning":"deontological","policy":"liability","emotion":"outrage"}
]