A new wave of disturbing AI-generated videos is testing YouTube’s ability to keep child-targeted content from sliding into violent, sexualized or abusive territory. WIRED found dozens of channels using bright animation, familiar children’s characters, nursery-like audio and searchable tags while showing scenes that include gore, cruelty and body-horror transformations.
The pattern echoes Elsagate, the 2017 controversy in which children’s characters appeared in dangerous, sexual and abusive scenarios. This time, generative AI appears to be making the problem faster to produce, easier to repeat and harder to stamp out channel by channel.
What WIRED found on YouTube
One channel highlighted by WIRED, Go Cat, presents itself as entertainment for children. Its description says, “Welcome to Go Cat—a fun and exciting YouTube channel for kids!” and claims that each episode is made for children to enjoy. The channel had 24,500 subscribers and more than 7 million viewers.
But the videos described by WIRED are not simple cartoons. In one example, a minion-like character falls into radioactive slime, mutates into a monstrous form, and attacks a screaming child. Other videos across related channels use familiar or kid-coded imagery while leaning into gore, abuse, medical horror and fetish-like themes.
WIRED said it found such channels through searches for ordinary terms including “minions,” “Thomas the Tank Engine,” and “cute cats.” The videos included long-running Elsagate-style material involving Elsa and Anna, as well as minions, animated cats and kittens.
The cat videos were described as especially prominent. Some feature titles such as “Kitten abused by its own mother” and show kittens being starved, forced into chores, beaten by parents, hospitalized and revived. WIRED also found channel names that were nearly alike, including “Cute cat AI” and “Cute cat of Ni.”
Why AI changes the scale of the problem
Elsagate content in 2017 often relied on traditional animation or actors in costume. The new wave still includes some of those methods, but generative AI changes the economics and speed. A creator no longer needs the same production skill, budget or time to make large quantities of animated content.
That matters because YouTube’s enforcement has to deal not only with single videos, but with systems of near-duplicate channels and reposts. WIRED reported that when one channel is flagged and removed, another channel carrying identical reposts can appear days later.
Robbie Torney, senior director of AI programs at Common Sense Media, told WIRED: “This trend is particularly concerning because of the scale and speed at which AI can generate this content.” He added: “Unlike traditional content creation, AI-generated videos can be produced in large volumes with minimal oversight. Without human review in the creation pipeline, inappropriate and potentially harmful material can easily reach kids.”
Common Sense Media reviewed several channels found during the investigation. The organization identified recurring themes including “characters in extreme distress or peril,” “mutilation, medical procedures, and cruel experiments,” and “depictions of child abuse and torture.”
How the videos signal to young audiences
Many of the channels do not need to say directly that they are made for children to create that impression. WIRED described videos using babies’ laughter, babbling, music and bright Cocomelon-esque backgrounds. In some cases, Cocomelon itself appears in the background.
Channel descriptions and metadata also matter. Go Cat openly advertised itself to children, while other channels claimed to be “not for kids” or avoided naming an audience. WIRED found videos tagged with terms such as #funnycat, #familyfun, and #disneyanimatedmovies.
Other videos involving polar bears and reindeer infected with parasites used tags such as #animalrescue. That suggests the content may be trying to appear near educational or family-oriented videos, even when the actual scenes are disturbing.
The result is a mismatch between packaging and substance. A young viewer, or a recommendation system following familiar keywords, may encounter material that looks playful from the outside but contains images of injury, abuse, fear and transformation once played.
YouTube’s response and what remains unresolved
YouTube told WIRED it “terminated two flagged channels for violating our Terms of Service” and was suspending monetization for three other channels. A YouTube spokesperson also said, “A number of videos have also been removed for violating our Child Safety policy.”
The spokesperson added: “As always, all content uploaded to YouTube is subject to our Community Guidelines and quality principles for kids—regardless of how it’s generated.” When asked about banned users opening new channels, YouTube said that would violate its Terms of Service and that enforcement uses “a combination of both people and technology.”
WIRED confirmed that some flagged channels had been removed, including two cat-centric channels featuring abuse themes. Yet it also found that other linked channels with reposts of the same videos remained online. Go Cat was still active, and its description had not changed.
YouTube’s earlier response to Elsagate was large. The platform removed ads on over 2 million videos, deleted more than 150,000, and terminated 270 accounts. WIRED reported that later changes, including new rules in 2019 tied to the US Children's Online Privacy Protection Act, mean channels of this kind now generally appear on YouTube’s main app rather than YouTube Kids.
The bigger moderation challenge
The core issue is not only whether a single video violates a rule. It is whether platforms can detect a production pattern built around cheap AI animation, familiar characters, child-friendly cues and rapid reposting.
For families, the investigation shows why surface signals are unreliable. A channel name, a cute animal, a familiar cartoon background or a family-friendly hashtag may not reflect what a video actually contains.
For YouTube, the challenge is more structural. AI-generated cartoons can be made quickly, copied easily and framed with metadata that helps them travel through search and recommendations. Even after removals, WIRED found that parts of the network remained visible.
The second wave of Elsagate, as described by WIRED, is therefore not just a replay of an old platform scandal. It is the same incentive problem under faster conditions: cheap content, childlike packaging, attention-driven distribution and harmful material that can reappear almost as quickly as it is removed.