At a time when more and more young children are hooked on digital devices, YouTube is giving them a jolt of AI.
After researching over 1,000 YouTube shorts that the video platform recommended to young children. of new york times We found that this algorithm significantly promoted AI-generated content that explicitly targeted “toddlers” and “preschoolers.”
In addition to being nonsense, videos are often presented under the guise of being educational. Two common themes are teaching children about the alphabet and animals. These themes provide simple structures that can be easily created with low effort.
Calling this video educational is also a stretch. 1 video highlighted in new york times It shows how a sticky liquid is squeezed into a glass of water and then transformed into different animals representing each letter of the alphabet. Only the animals are strange chimeras with mermaid tails. In another set, set to an off-key rendition of “Old MacDonald Had a Farm,” a giant egg tumbles through a barn door before hatching into an impossibly proportioned horse. And in another short alphabet, a quail transforms into an aerial drone and a rhinoceros transforms into a dump truck with a giant animal’s head on it.
At best, these videos are just redundant regurgitations of mindless “Cocomelon” style content. But in the worst cases, experts worry they could actively harm their cognitive development.
“For me, the pointlessness of these videos is a big problem because they’re just trying to get attention,” said Jenny Radesky, a developmental behavioral pediatrician and associate professor of pediatrics at the University of Michigan Medical School. new york times. “And the worst case scenario is that it’s so fantastical and attention-seeking that it overloads the child’s cognitive abilities.”
Surrealistic visuals used in many AI videos — some examples highlighted in new york times is as follows — Radesky speculated that it may interfere with young children’s ability to distinguish between fantasy and reality.
It’s not a niche problem. YouTube’s algorithm seems to be surprisingly enthusiastic about recommending AI slop. In that test, new york times First I watched a popular children’s channel, then scrolled through short videos.
More than 40% of the videos that followed during the 15-minute session appeared to include AI visuals. This is amazing. Rather than recommending traditional children’s content, the algorithm appears to have gravitated toward AI by default.
“When I watched channels like ‘Ms. Rachel’ and ‘Bluey,’ I expected to see more content along those lines, more ‘Bluey’ shorts,” he said. new york times Ariheta Laika, the reporter who led the investigation, spoke in an interview on the paper’s Hard Fork podcast. “And I didn’t really see it.”
While not all parents have the same concerns about AI-generated images, experts say it’s hard to deny what this technology is good at and what it’s generally actually used for: quickly creating short-form, often absurd content with no plot or message — the opposite of what kids should be seeing. Rachel Barr, developmental psychologist and director of Georgetown University’s Early Learning Project, said: new york times Rather, children learn best from media that has clear narratives, characters and scenes that relate to real life.
For example, there are few relatable elements from real life that a child can glean from a surreal clip of an animal jumping off a diving board. In theory, someone could use AI to create thoughtful educational videos for kids, but that’s not why YouTube Shorts are proliferating and rapidly gaining views.
“At least when you’re watching a normal anime, there may be moments of relative calm, or the story unfolds over the course of a few minutes,” journalist Casey Newton said, sounding moderately perturbed during an interview with Laika at Hard Fork. “If you’re just showing a child raw visual stimuli and bombarding them with it, it’s probably not going to be all that good for them.”
Newton speculated that slop makers preferred the alphabet because the subject matter was well-suited to stringing together a bunch of short clips while maintaining “a kind of coherence.”
For now, the long-term effects of watching addictive AI short videos are unknown, even if the videos seem shamelessly designed to be as addictive and mind-numbing as possible. “These really seem like they’re designed to get inside your head,” said Mitch Prinstein, a professor of psychology and neuroscience at the University of North Carolina. new york times. “It could even be harmful, but we need more data.”
However, other research suggests that other forms of AI use, such as reliance on chatbots, can impact cognitive skills such as critical thinking, even in adults. Additionally, more research is being done on how exposing children to “brain rotting” content can have negative effects, and there may be a link between screen time and ADHD diagnosis, for example.
YouTube requires creators to disclose whether they used AI to create “realistic content,” but does not have an AI label requirement for the cartoonish style used in short videos aimed at children. That leaves the burden on parents to closely monitor what their children are viewing in apps that offer infinite scrolling of content. The most sensible and practical solution is to not let children use these apps at all, but this ignores the reality that many parents rely on digital devices to entertain their children.
Learn more about AI: ‘Educational’ YouTube AI slop encourages kids to play in traffic
