‘AI slob’ is changing social media… And the growing backlash
-
- reporter, Joe Tidy
- reporter, Cyber Correspondent, BBC World Service
Theodor clearly remembers ‘Slob’, the artificial intelligence (AI) that made him behave strangely.
Two very skinny boys, who appear to be Asian, are staring at the screen with a birthday cake in front of them in the middle of a road in heavy rain.
For some reason, although his face is clearly a boy, he has a thick beard on his face. One child had both hands amputated, and another child said, ‘Today is my birthday.’ He is holding a placard with the phrase ‘Please click ‘Like’.’
Although the image was full of clues showing that it was an AI creation, the content went viral on Facebook, receiving nearly 100 likes and heart emoticons.
Theodore, looking at this, felt that he could no longer bear it.
“I felt like my heart was being crushed,” the 20-year-old from Paris, France recalled. “Absurd AI-generated images were flooding Facebook, and they were getting this much attention without any verification. It felt really crazy.”
So, he created an account called ‘Insane AI Slob’ on Theodore defines ‘AI slop’ as low-quality, false, and unconvincing content like filth that is quickly created with AI.
His account attracted attention, and soon he was flooded with reports about popular AI slop content.
image copyright, Left: Théodore. Right: “Baby Bubbles and Babbles” Facebook page.
There were similarities among the reported contents. Many of them were related to religion or the military, or showed poor children doing touching things.
“Content about children in third world countries doing touching things is always popular,” he said. “For example, there is a scene where poor children in Africa create enormous statues out of trash. People think this is cool, so the creators seem to think, ‘Okay, I should make more content like this.’”
The number of followers on Theodore’s account quickly exceeded 133,000.
AI slop is rapidly encroaching on social media and is currently unstoppable. Technology companies have embraced AI. Some companies have announced that they will crack down on some AI ‘slobs’, but many social media feeds are still full of this filth.
In just a few years, the experience of using social media has changed significantly. So how did this happen, and what impact will it have on our society?
And perhaps the most pressing question is: how much do the billions of social media users around the world care about this issue?
‘The third stage’ of SNS
Table of Contents
Last October, at another performance presentation held in an excited atmosphere, Meta CEO Mark Zuckerberg happily declared that social media had now entered its third phase centered on AI.
“The first stage is when all of your content comes from friends, family, and accounts you directly follow.”
He then announced to investors, “The second phase is the addition of content creators. And now that AI makes creating and recombining content easier, we will be adding another vast pool of content.”
Meta, which operates several social media platforms such as Facebook, Instagram, and Thread, not only allows users to post AI-generated content, but has also released tools to assist in such creation. In addition to tools to help create images and videos, increasingly powerful filters are available across the board.
When asked for an opinion on this, Meta referred to the BBC’s earnings announcement in January of this year. At the time, the CEO said that Meta was focusing more on AI. However, there was no mention of strengthening AI slop regulations.
At the time, CEO Zuckerberg predicted, “Soon, you will witness explosive growth in new, more immersive and interactive media formats made possible by advances in AI technology.”
In a blog post predicting 2026, YouTube CEO Neil Mohan revealed that in December of last year alone, more than 1 million YouTube channels produced content using the company’s AI tools.
CEO Mohan explained, “Just as synthesizers, Photoshop, and computer graphics have revolutionized the field of sound and visual effects, AI will be useful to producers who want to actively utilize it.”
Acknowledging growing concerns about “low-quality content, aka AI slop,” the company added that it is working to improve its systems for detecting and removing “low-quality, repetitive content.”
However, he also did not judge which content should and should not be allowed to trend. However, he mentioned that ASMR and live game play broadcasts, which were niche content in the past, have now become mainstream content.
Meanwhile, according to a study by AI company ‘Capwing’, 20% of all content currently exposed to newly opened YouTube accounts is filled with “low-quality AI videos.”
This phenomenon was especially noticeable in short-form videos (video content with a short playback time). Low-quality AI content was found in 104 of the 500 YouTube shorts (YouTube’s short-form videos) that were first exposed to the newly created account by the researchers.
The creator economy, in which creators and channels make money from participation and views, appears to be the main cause of this phenomenon. If you look at the number of views of some AI channels and videos, you can see that users are actually immersed in such content. Or it could be the result of an algorithm that determines what we see.
According to Capwing, the AI slop channel with the most views is India’s ‘Bandar Apna Dost’, with 2.07 billion views, and its creators are estimated to earn about $4 million (about 5.7 billion won) per year.
However, there is also quite a bit of backlash against this AI slop.
In the comments section of many AI video or image posts that have gone viral, it is now common to see angry reactions criticizing the content.
Giant monsters and hideous abdominal parasites
Theodore helped lead this backlash.
He used his newfound influence on He claimed that some of the videos were offensive and harmful, and that some even appeared to be aimed at children.
These videos feature gory visual images with titles such as ‘Mother cat saves kittens from horrific abdominal parasite’.
There is also a video of a woman in pajamas who is transformed into a huge monster after eating parasites and is eventually healed by Jesus.
image copyright, ‘Sprung Nexus’ YouTube account
YouTube explained that it had deleted the channels that uploaded the videos, saying it was a violation of its community guidelines. They added that they are “focused on ensuring users have access to high-quality content, regardless of how it was created,” and that they are working to “reduce the proliferation of low-quality AI content.”
However, as this experience repeats itself, Theodore is exhausted.
Even lifestyle platforms that seem relatively cozy, such as Pinterest, where recipes and interior design ideas are shared, cannot escape this influence.
Users became outraged by the flood of AI slop, so Pinterest introduced an opt-out system that can block AI-generated content, but for this to work properly, users must acknowledge that the images of their perfect homes are AI-generated.
Comment window full of anger
In the feed (of course, I know that each person’s feed and comments are different), there is a constant backlash against AI slop.
Wherever you go, including TikTok, Threads, Instagram, and X, there are people opposing such content.
Sometimes, comments criticizing an AI creation receive more likes than the original post. This is the case in a recent video depicting a snowboarder rescuing a wolf from a bear. The video itself received 932 likes, but the comment “Who’s sick of this AI garbage?” received 2,400 likes.
But of course, all of this ends up raising a monster.
For SNS platforms, where the key is to keep users from scrolling, all forms of user participation are ultimately good participation.
So does it really matter whether the surprising, touching, or shocking videos that appear on our social media feeds are real or not?
‘Brain rot’ effect
In response to this question, Assistant Professor Emily Towson, who studies political science and disinformation at Syracuse University, said that it depends on what people do on social media.
Professor Towson added, “If you are someone who enjoys a short-form platform simply for entertainment purposes, the standard for judging the value of content is simply ‘Is it entertainment?'” He added, “However, if you are someone who uses the platform to learn about a topic or communicate with community members, you may perceive AI-generated content as a problem.”
Additionally, users’ feelings about AI slop also vary depending on how it is communicated.
If the content is made to make anyone laugh, it seems to be taken as a joke. However, slop that is designed to deceive viewers can make people angry.
An AI-generated video I recently came across shows this well. It was an incredibly realistic animal documentary-style video that featured amazing hunting scenes of leopards. And looking at the comment window, some viewers believed it was a real video, while others were suspicious that it was an AI creation.
Another comment that caught my eye was, “Which documentary is this scene from? Please tell me. It’s the only way to prove that it’s not an AI (creation).”
Alessandro Galeazzi, assistant professor at the University of Padova in Italy, studies SNS behavior and echo chamber effects.
Professor Galleacci pointed out that mental effort is required to confirm whether a video is an AI creation, and that he is concerned that people will eventually give up on confirmation in the long run.
In addition, the explanation is, “People’s concentration may be weakened by the flood of ridiculously low-quality content created with AI.”
He warned that obviously fake AI slop, like a fish wearing shoes or a gorilla working out at the gym, is different from content intentionally created to deceive viewers, but even such content can have harmful effects. This is the risk of the so-called ‘brain rot’ effect. I believe that continuous exposure to SNS is damaging our intellectual abilities.
“AI slop accelerates the brain rot effect,” he said, “because it makes people quickly consume a lot of content, even though they know that it is not only untrue, but has no meaning or interest.”
Reduce content review staff
Meanwhile, beyond simple low-quality AI slop, some AI-generated content can cause much more serious problems.
Elon Musk’s AI company xAI and social media platform
And after the US attack on Venezuela, a fake video was spread showing Venezuelan citizens crying on the streets and expressing gratitude to the US. Such content can shape public opinion in a certain direction and give the impression that U.S. airstrikes received greater public support than they actually did.
In response to this, experts emphasized that it is particularly concerning because so many people only get news through social media.
Manny Ahmed, CEO of Open Origins, a company that differentiates between AI creations and real images, says that for those who create content that is not an AI creation, a new way to prove that the content is real is needed.
CEO Ahmed emphasized, “We have now reached a stage where we cannot be sure what is real just by looking closely,” and added, “Instead of finding out what is fake, an environment should be created where real content can publicly prove its source.”
You may be thinking that social media companies can take on this task. However, many social media companies, such as Meta and
SNS without slop?
If most existing technology giants are practically happily ignoring the influx of AI slop, will they be able to challenge their stronghold if a new SNS platform that promotes a slop-free environment emerges?
But the possibility seems small. This is because detecting AI products is becoming increasingly difficult. Machines can no longer accurately detect whether a video or image is clearly fake, and whether or not content is ‘slop’ is a subjective judgment, so machines will have their limitations.
However, if a new SNS appears and people turn their feet (or, more accurately, their fingers), meaningful changes may occur.
In the case of the French app ‘Breal’, which became popular during the COVID-19 pandemic, users are asked to take selfies of their true selves without filters at randomly selected times.
Breal has not yet grown into a huge platform like Facebook or Snapchat, and it probably won’t in the future. But it has succeeded in making other platforms nervous, and some have even copied the idea.
If a new challenger platform emerges to compete with AI slop, this phenomenon may repeat itself.
Meanwhile, in Theodore’s case, he feels that he has already lost this battle, and predicts that AI slop will never go away.
Although he still receives daily reports from his 130,000 followers, he no longer posts frequently and seems somewhat resigned to the online environment that has now become the new normal.
“Unlike many followers, I am not unconditionally opposed to AI itself,” he concluded, “I am simply opposed to online pollution caused by AI slop created for quick fun and views.”
Top image source: BBC / AI image created with Adobe Firefly
