Posted on 12/28/2025 3:41:22 PM PST by E. Pluribus Unum
Kapwing reports up to 33% of YouTube feeds are low-quality AI content. Spain and South Korea top in subscribers and views, sparking global concern.
Once a fringe term, "AI slop" has become part of the global vocabulary. In 2025, both The Economist and Australia’s Macquarie Dictionary highlighted the phrase as shorthand for the surge of low-quality, auto-generated content flooding the internet — from shallow articles and spammy videos to algorithmic filler produced at scale. That growing frustration has data to back it up.
A new report by Kapwing suggests that between 21% and 33% of YouTube feeds, as of October 2025, may consist of AI slop or closely related "brainrot" videos, underlining how rapidly this form of content has spread across platforms worldwide.
While AI-generated video has found legitimate applications in filmmaking, advertising and education, with film schools and major brands experimenting with the technology, the report suggests that a parallel ecosystem has emerged. In this space, creators use prompt-based tools to mass-produce content at scale, flooding platforms with repetitive, low-quality visuals that prioritise engagement over originality.
Kapwing defines AI slop as careless, low-quality content generated through automated applications and distributed to attract views or influence opinion. Brainrot, a closely related category, refers to compulsive, nonsensical videos that corrode attention and comprehension through endless repetition and spectacle.
To measure how widespread this phenomenon has become, Kapwing examined the top 100 trending YouTube channels in every country, identifying which among them primarily produced AI-generated slop.
The researchers then analysed view counts, subscriber numbers and estimated revenue using data from Social Blade. To simulate the experience of a new user, they also created a fresh YouTube account and tracked the first 500 Shorts that appeared in the feed.
The findings reveal striking geographical patterns.
Spain emerged as the country with the largest combined subscriber base for trending AI...
(Excerpt) Read more at cnbctv18.com ...
I am thinking folks are using the comments in the Facebook and Youtube posts to train/correct that AI slop. For example - “why did that truck have one two tires?” That way they can use the commenters to improve.
I am starting to think we should poison pill these comments. Say stuff like “I knew it was AI because all the people had five fingers on their hands.” Or “can you believe that car had four tires? Of course that was AI.”
Who’s with me on this project? Get them so messed up that they can’t trust the in
put.
https://www.dailymail.co.uk/news/article-15367387/texas-town-furious-bitcoin-mara-holdings.html
AI uses much the same. If everything is going to be AI, they need to have these companies coming with their own nuke power plant and desalination plants.
>> Yes, many are good, and that’s what’s scary.
Agree.
I know someone who enjoys conversing with one of those AI Bots (”Grok” I believe). Now they have not only his likeness but his voice and speech patterns as well. High risk if you ask me!
Agreed. It’s for reasons like this that I don’t have pictures on my cell phone or any social accounts. It’s just too dangerous now.
AI will destroy society as we know it and bring about unemployment and poverty which will fuel social unrest and overthrow of Governments.
Thanks...
Still can’t get my head around it
.
But I’ll keep Digging.
What?
Wait
You can’t say That on TV.
That’s a Great Video
By the way
People are developing along with the technology in their ability to recognize fakes. After you experience a few of these you are able to recognize them for what they are.
People have always been able to propagandize and lie to us, there have always been scams, urban legends, etc., and we’ve had to learn to see through it all. This is just a new way of doing it.
The only real control I can see lies in learning to question and use one’s common sense.
I think it’s late for that. They’ve probably been training them on Internet content for years, if not decades. (Some of the very early chatbots were on line precisely so that people could chat with them and the program would learn from the interaction.)
Ping to look at ....
again and again and again.
Almost all top 10 videos are AI now. I am getting better at spotting some before I waste my time and click.
These are obviously AI and are entertaining. I hate the other ones, where you click on it first to find out they are AI.
Have you seen the bigfoot videos or the Star Wars ones?
The giveaway is when they quote years, like 1945, as “One thousand, nine hundred, and forty-five”.
Yea, so much WW2 content is AI crap now. I only watch the ones that are older now. I have been sucked in by some new posters and all are AI crap with the same announcer. The minute I hear his voice, I am out.
All you need is Indy Neidell’s Time Ghost presentations.
That and release date, but yeah, we’re on the same page.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.