Why TikTok’s AI-generated cat could be the future of the internet

Tearjerker videos of AI-generated cats earned millions of views and a devoted following, blurring the line between spam and art. Is it the algorithm, or is this what the internet wants?

There have been many famous internet cats, but it’s possible no internet cat has made more people cry than Chubby. Depictions of Chubby vary, but he is always rotund, ginger and AI-generated. He is almost always involved in a deeply sad or peculiar situation. And he has baffled, outraged and won over millions of people. 

Content creators on TikTok and YouTube Shorts tell stories about Chubby and his family in wordless slideshows of AI-generated pictures. A recent video by the TikTok account @mpminds opens with Chubby and his child, Chubby Jr, dressed in tatters. Chubby holds a cardboard sign that reads “Will Purr Fro Eood” (AI image generators can churn out impressive graphics, but they’re notoriously bad at rendering text). In the next images we see Chubby shoplifting from a grocery store, getting arrested by the police and leaving a distraught Chubby Jr to an uncertain fate. The last image shows Chubby behind prison bars, dreaming wistfully of his son. The video has over 50 million views and 68,000 comments, written in several different languages. 

Chubby is far from alone. Similar content has flooded social media in recent months. In March, reports emerged that bizarre AI-generated images of Jesus made out of shrimp exploded on Facebook, baffling observers and racking up millions of views. The posts were so popular that many speculated the social media engagement was coming from networks of bots programed to pose as humans. But with this new genre of cat videos, it’s clear the internet has turned a corner. There’s no question that Chubby and his AI-generated feline friends are capturing real people’s attention and emotions, en masse. It’s raising new questions about the meeting of art and technology, and perhaps offering a vision of the internet’s future.

Most videos place the cats in depressing human situations. Chubby Jr faces schoolyard bullying. Chubby is addicted to cigarettes. The cats get drafted and go to war. And for the past three months, the videos have almost all been soundtracked by an AI-generated cover of Billie Eilish’s What Was I Made For, with the lyrics swapped out for meows. Before that, the standard song was a meowed AI cover of Sia’s Unstoppable. 

“can’t let gang know i teared up to this” reads one typical comment on a TikTok by @relatablecutecats (160,000 followers) about Chubby Jr failing a test at school. “Out of all things happening in this world this is what I get sad at,” comments another viewer on a video about Chubby Jr getting abducted by a pigeon while eating McDonald’s with his father.

The AI-generated cat stories are objectively weird. They are also wildly popular, and more than a little controversial. But whether or not Chubby is what people are looking for when they go online, this is the content the internet is giving us – at least for now. The only question is why?

Whether they’re grumpy, keyboard playing or “Nyanning”, cats have always been at the heart of digital culture. Jessica Maddox, professor at the University of Alabama and author of The Internet Is for Cats: How Animal Images Shape Our Digital Lives, says cat content is a hit not just because it’s cute, but because “images of cats can be malleable… we can make cat images mean whatever we want them to mean”. In the Victorian era, Maddox says, “people wrote letters to each other in their cats’ voices and printed off photo-plates of their cats to share with friends”.

Cats were a natural fit as meme culture took root online. And as technology advances, pet lovers have moved from photos of cats to generating them with artificial intelligence. The debut of free-to-use and widely available AI generators like Midjourney, ElevenLabs and DALL-E has changed the look, feel and landscape of the web. By putting powerful tools into the hands of anybody with an internet connection and a sufficient amount of chutzpah, these AI generators catalysed a rapid bloom of new genres of online content.

“I began in January 2024”, says Charles, the creator behind the popular @mpminds account who asked to withhold his full name to protect his professional reputation. “I saw another account which made AI-generated pictures of cats, not the same content, but kind of the same vibe. I remixed it to make stories, and created the characters of Chubby and Chubby Jr. So it was maybe sketched out a bit before me, but I took the road and customised things to make it what it is today.”

Many creators have followed the same path, using AI tools to remix existing culture the way other internet users riff on movies, music, politics and all else. “I knew there was potential to make money on TikTok, and I saw that AI-generated content was quite popular,” says Charles, who works in the finance business in France. “I experimented with different ideas before the cat videos, and it was the cat videos that really took off, so I stuck with that.”

“A lot of people don’t realise how AI has become so baked into social media, both in terms of features and shareable content. Someone might be sharing a popular social media post that is AI without realising it”, Maddox says. But as more and more AI-generated content appears online, there is also more and more “pushback”, she says.

Users and researchers have noted an increase in what’s come to be known as “AI slop”, low-quality content manufactured in large amounts by people using AI generators. The most famous example is probably the “Shrimp Jesus” study from early 2024, in which researchers at Stanford and Georgetown documented networks of AI-generated spam accounts on Facebook. These accounts posted surreal AI-generated images dozens of times a day, garnering hundreds of millions of likes and views. One AI-generated post was reportedly among the top ten most-viewed posts on the entire platform in the third quarter of 2023.

By enrolling in the monetisation programs of the platforms or directing viewers to external links and services, users creating this kind of content can make money. There’s also money to be made in teaching others to generate content with AI. As reported in the Washington Post, AI cat stories creators sell courses that instruct students in how best to gain a following and make money. 

“AI made it possible to do this at scale, so why not try?” says Renée DiResta, the former research director at Stanford who ran the Shrimp Jesus study. The volume and speed that AI generators allow spammers to post at means they have more shots at the moving target of virality. According to DiResta, spammers and scammers using generative AI also take advantage of social media algorithms which are better at recommending content than evaluating its accuracy or relevance. In 2023, at the same time as AI-generated tools enabled spammers to easily produce high volumes of content, Facebook reportedly tinkered its algorithms so they served users more content from accounts they had not previously interacted with. Paired with an inadequate system for labelling AI-generated content, this created a perfect storm.

One way to read the rise of AI-generated content is to see it as a consequence of social media algorithms themselves. Algorithms like Facebook’s or TikTok’s filter billions of data points per day and evaluate that content through automated programs (many of which involve AI). In addition to managing the flows of content across vast networks, the preferences of these algorithms actively shape those flows, determining what goes viral and what doesn’t. AI-generated content can succeed by strategically playing to the preferences of algorithms first and human audiences second. “Machines are making content for machines,” DiResta says.

But the machines still require some creative input from real people. “I believe AI-generated content can be art,” Charles says. An AI image generator may be doing the bulk of the work compared to the efforts of a human illustrator, but it’s on creators like Charles to come up with ideas and themes, and he says it still takes about an hour to make just one cat story video. And by any metric, Charles has been dramatically successful, both in terms of likes and views and the emotional response of his audience. “The person generating the AI content has a significant role in the outcome,” he says.

“My initial reaction was mostly, ‘what is this? Why is this?'” says Daniel Chartier, a painter who first encountered the AI cat stories on his TikTok For You page. “I kinda went through the five stages of grief and then finally landed on acceptance. It upsets me how actually sad I find them.”

For Chartier, whose art often focuses on animals, imperfection is a key part of the AI cat stories aesthetic. Chubby and the other cats are rarely consistent between images: costumes and fur patterns change, backgrounds alter dramatically. But despite the inconsistencies and occasional hallucinations of the AI generators, he says the cat stories remain compelling because of how “devastating” they are. “I just love that I can invest in these characters while everything else is so goofy. It seems like it would not be emotional, the way you could create anything that looks like this, but it’s successful sometimes. The contrast is great,” Chartier says.

Charles agrees. “A good cat video is one with a tragic ending, a devastating one. The reason is, firstly, because I wanted to amuse myself; to put the cats in crazy situations. I also saw that sad videos generate more engagement because they elicit the compassion of viewers.”

The whiplash of feeling a genuine emotional response from a clearly AI-generated video seems to be one of the main reasons people want to watch bad things happen to Chubby. In the comments on many AI cat stories, users report being brought to the videos by another even more viral video in June by TikTok user @b.ajasiii which shows a toddler watching AI-generated cat videos and weeping. 

The video received 173 million views and a repost by Billie Eilish, whose song is used in many AI cat stories. And it sparked a trend of other TikTokers filming their toddlers’ reactions to the cat stories. “That was a key moment (for the AI cats trend),” Charles says. “I was already viral before that. But (that video) made my account explode even more.”

But many aren’t thrilled with the trend. A creator who goes by the name Tommy Guacamole was one among many who made a video mocking people’s responses to the cat videos. It’s funny how easily these cat posts are able to manipulate our emotions, but “I think AI content is trash and genuinely ruining the internet”, he says.

The AI cat stories are not trying to hide the fact that they are made using AI. TikTok and YouTube both require creators to label AI-generated content as such, and the cat content accounts usually comply with this rule. Several even brand themselves as AI accounts in their usernames and bios. But the fact that they are AI-generated does not seem to stand in the way of the audience’s heart strings.

“I’m not sure where the project is going,” says Charles, “For now, I’m going to keep posting the content as it is and see what happens.”

For timely, trusted tech news from global correspondents to your inbox, sign up to the Tech Decoded newsletter, while The Essential List delivers a handpicked selection of features and insights twice a week.

For more science, technology, environment and health stories from the BBC, follow us on Facebook and X.