Although AI is still in its infancy, we’re already seeing some impressive tools coming out which have also caused quite a bit of concern as well. While tools from Google are always welcome, especially when they can help us to learn in new ways, the brand has also introduced some lackluster and problematic products as well.
However, its latest move has struck quite a nerve, with the BBC reporting that YouTube has been using AI to alter the look of content on the platform. Now, this in itself would be quite surprising, but it goes even a step further with YouTube not even telling the original creators that the alterations were taking place, or even asking permission to get this done.
YouTube might want to roll this change back
Why this is gaining so much attention now is that a major creator, Rick Beato, who is known for in-depth videos on everything music, noticed the changes with the content on his own channel. Beato goes on to explain that he saw that his hair looked ‘strange’ and in other instances it looked like he ‘was wearing makeup’ in some of his Shorts.
The BBC states that YouTube has been using AI in order to edit videos on the platform to enhance their appearances. The problem is that most of these changes don’t even look good, and it adds a quality that really isn’t pleasing to the eye. If you’ve ever watched an AI-generated video, you probably know what we’re talking about here.
It’s when AI can’t quite deal with all the intricacies that are in a photo or video, and decides to create its own twisted image of what it thinks things should look like. It’s quite fascinating to see all of this unfold, especially with people’s apprehension about AI. If you want to see the full breakdown of this ‘AI magic’ we highly suggest checking out this video by Rhett Shull, which we have linked to in this article.
Shull does a deep dive into what he’s found, showing that while the changes could be very minor in some cases, it’s still being done without permission from the creator. Of course, there are many reasons why creators might not agree with this particular use of AI, but as the AI gets better, it’s going to become harder to distinguish between what’s real and what’s not.
YouTube via its Liaison confirmed on X that it is indeed running some kind of ‘experiment’ that uses ‘traditional machine learning technology’ to ‘unblur, denoise, and improve clarity’ of videos. While it’s unclear just how many videos have had this type of alteration performed, it appears that this type of activity might have been going on for some months now.
Judging from the responses to the post on X, it might have been a good idea to announce this change before it even got started. It’s unclear how things will progress, but it’s quite possible YouTube will revert things back to normal since it’s facing so much public backlash. We’ve been watching content just how the creators intended for years now, so why not just keep these original creations as is?
While creators deserve to feel safe publishing their work to YouTube, creators are bound by a set of terms when they sign up, which mostly likely allows this type of thing to happen. We’re not saying it’s right, it’s just that, with so little control, this probably isn’t going to be the last time that something like this happens.
Plus, once YouTube gets the tech down, and it’s so good that we can’t even distinguish the difference, how are we to even know? All we can hope for is transparency in these types of situations. We’ve seen plenty of missteps on this new and untravelled road. And as mentioned before, AI is still in its infancy, and there’s no telling how the world will look in just a couple of years because of how far and fast things are progressing.
