How old is the strange kid singing on youtube

And there is a lot of movement; sometimes every pixel of the screen seems to be in motion. Krishnan and Chandar believe that any given shot needs to include many different things a child could notice: A bird flying in the background.

Something wiggling. The men know this with quantitative precision. If a video achieves a 60 percent average completion rate, ChuChu knows it has a hit. But what people want changes.

ChuChu learns many lessons from parents, who provide the company with constant feedback. It heard from parents who questioned the diversity of its characters, who were all light-skinned; it now has two light-skinned and two dark-skinned main characters.

It heard from parents who wondered about the toy guns in one video; it removed them. ChuChu is largely making things up as it goes, responding—as any young company would—to what its consumers want. Part of the absurdity of the internet is that these questions get asked only after something metastasizes and spreads across the world. It created an unprecedented thing— Sesame Street —with help from a bevy of education experts and Jim Henson, the creator of the Muppets.

The cast was integrated. The setting was urban. The show was ultimately broadcast on public television across America, defining a multicultural ideal at a time of racial strife. The s and s saw the growth of cable TV channels targeted at children. With the rise of ubiquitous merchandising deals and niche content, powerful American media companies such as Disney, Turner, and Viacom figured out how to make money off young kids.

Since then, however, little kids have watched less and less television; as of last spring, ratings in were down a full 20 percent from just last year. As analysts like to put it, the industry is in free fall. The cause is obvious: More and more kids are watching videos online. This might not exactly seem like a tragedy. After all, Americans watch a lot of TV. By the time Nielsen began recording how much time Americans spent in front of TV screens in —50, each household was already averaging four hours and 35 minutes a day.

That number kept going up, passing six hours in —71, seven hours in —84, all the way up to eight hours in — Viewing finally peaked at eight hours and 55 minutes in — Considered purely as a medium, television seems to have little to recommend it over YouTube.

The institutions of the 20th century shaped television into a tool for learning. At first, pretty much everybody agrees, television for kids was bad—dumb cartoons, cowboy shows, locally produced slop. Their shallowness of thought and feeling is markedly apparent, and they display a lack of cooperation and inability to finish a task.

But not much happened, and the government and TV networks generally settled into a cycle that has been described by the media scholar Keisha Hoerrner. Absent substantive oversight by regulators, in the late s the calls for change entered a new, more creative phase. The Corporation for Public Broadcasting was formed in with government dollars. These shows were tremendously successful in creating genuinely educational television.

Another study found that regular adult TV stunted vocabulary development, while high-quality educational programs accelerated language acquisition.

The most fascinating study began in the s, when a University of Massachusetts at Amherst team installed video cameras in more than homes, and had those families and hundreds of others keep a written log of their media diet. The team was unequivocal about the meaning of these results: What kids watched was much more important than how much of it they watched.

S o what message are very young kids receiving from the most popular YouTube videos today? And how are those children being shaped by the videos? A crowd waves its hands in the foreground. Lights flash and stars spin in the background. Johnson told me all that movement risks distracting kids from any educational work the videos might do. For kids to have the best chance of learning from a video, Johnson told me, it must unfold slowly, the way a book does when it is read to a child.

Children under 2 struggle to translate the world of the screen to the one they see around them, with all its complexity and three-dimensionality. Most important for kids under 2 is rich interaction with humans and their actual environments.

Older toddlers are the ones who can get something truly educational from videos, as opposed to just entertainment and the killing of time. If kids watch a lot of fast-paced videos, they come to expect that that is how videos should work, which could make other educational videos less compelling and effective. ChuChu has changed over time—it has slowed the pacing of its videos, focused on the key elements of scenes, and made more explicitly educational videos. But in the wilds of YouTube, the videos with the most views, not the most educational value, are the ones that rise to the top.

Not to put too fine a point on it, but this is almost precisely the problem that the rest of the media world finds itself in. Because quality is hard to measure, the numbers that exist are the ones that describe attention, not effect: views, watch time, completion rate, subscribers. YouTube uses those metrics, ostensibly objectively, when it recommends videos. Some were sadistic or sick. Others seemed like grab bags of keywords that had been successful for more professional operations: nursery rhymes , surprise eggs , finger family , learning colors.

These were videos reverse engineered from whatever someone might enter into the YouTube search box. The world of YouTube is vastly different from the world of broadcast television. While broadcasters in the United States and abroad are bound by rules, and the threat of punishment for breaking those rules, far fewer such regulations apply to the creators of YouTube content, or to YouTube itself. In addition to its main site, however, the company has developed an app called YouTube Kids.

Like normal YouTube, it plays videos, but the design and content are specifically made for parents and children. YouTube Kids is not available in as many countries as normal YouTube is. Little kids are responsible for billions of views on YouTube—pretending otherwise is irresponsible. In a small study, a team of pediatricians at Einstein Medical Center, in Philadelphia, found that YouTube was popular among device-using children under the age of 2.

Oh, and 97 percent of the kids in the study had used a mobile device. By age 4, 75 percent of the children in the study had their own tablet, smartphone, or iPod. And that was in To date, YouTube has hidden behind a terms-of-service defense that its own data must tell it is toothless.

The company can declare its efforts for children sufficient at any point. But there is something the company could do immediately to improve the situation.

YouTube knows that I—and tens of millions of other people—have watched lots of videos made for toddlers, but it has never once recommended that I switch to YouTube Kids. Think of how hard Facebook works to push users from Instagram onto Facebook and vice versa. Why not try to get more families onto the YouTube Kids app? These promos have helped drive our growth. Today, YouTube Kids has over 14 million weekly viewers and over 70 billion views. If streaming video followed the broadcast model, YouTube—in partnership with governments around the world—could also subsidize research into creating educational content specifically for YouTube, and into how best to deliver it to children.

The company could invest in research to develop the best quantitative signals for educational programming, so it could recommend that programming to viewers its algorithm believes to be children. It could fund new educational programming, just as broadcasters have been required to do for decades.

Other, more intense measures could help, too. For example, how about restricting toddler videos to the YouTube Kids app? Toddler content could, in effect, be forbidden on the main platform.

We may earn a commission if you buy something from any affiliate links on our site. Learn more. YouTube videos using child-oriented search terms are evading the company's attempts to control them. In one cartoon, a woman with a Minnie Mouse head tumbles down an escalator before becoming trapped in its machinery, spurting blood, while her children baby Mickey and Minnie characters cry.

Mickey Mouse, Donald Duck Cartoon , racked up over three million views in a single day. It could be viewed even with YouTube's family-friendly restricted mode enabled and existed, along with plenty of similarly distressing content, on Simple Fun, a channel that had been in operation since July The channel has now been removed by YouTube "due to multiple or severe violations of YouTube's policy against spam, deceptive practices and misleading content or other Terms of Service violations.

WIRED found videos containing violence against child characters, age-inappropriate sexualisation, Paw Patrol characters attempting suicide and Peppa Pig being tricked into eating bacon.

These were discovered by following recommendations in YouTube's sidebar or simply allowing children's videos to autoplay, starting with legitimate content.

It might sometimes, but if it does it is coincidence," says former YouTube engineer Guillaume Chaslot, who founded AlgoTransparency , a project that aims to highlight and explain the impact of algorithms in determining what we see online. The company explained that it is increasing its efforts to control content that violates its terms and conditions. YouTube is home to millions of hours of children's entertainment — part of the hours of video uploaded to the service every minute — ranging from CBeebies and Disney to the incomprehensibly successful Little Baby Bum , a UK-based YouTube-native children's channel devoted to 3D animated songs and nursery rhymes for pre-schoolers in numerous languages.

Content for pre-school children, in particular, can be lucrative for ad-funded channels, as small children will readily watch and poke at whatever videos YouTube suggests, while harried parents are often unable to fully supervise every minute of their child's media consumption.

AlgoTransparency regularly indexes the kids' videos most likely to be recommended by YouTube. Its lists show that YouTube's most-suggested children's videos lean disproportionately towards a combination of YouTube-native songs and nursery rhymes designed for a US audience; long edited-together compilations of TV series such as Peppa Pig , and strange, low-budget 2D and 3D animated mash-ups of animals, characters and voice samples.

Previously described in James Bridle's Something is wrong on the internet Medium post as "decidedly off", the latter type of content can be loosely and collectively categorised as 'weird children's YouTube'.

Titles are typically a word salad designed to attract children's and parents' searches, while the videos' content leans heavily on generic 2D or 3D animated models, usually incongruously combined with familiar figures from hit Disney or superhero franchises. The sheer number of them on the platform is staggering, and many have millions or even hundreds of millions of views. Neither video contains any content more distressing than badly-animated video and the intensely annoying Finger Family song, but both are good examples of videos that use popular franchises and the promise of education to target searches that parents and children are likely to carry out.

Based on what YouTube insiders have said about how its suggestion engine works, racking up high view counts and minutes watched makes it all the more likely that videos will be recommended to others. Once you're into the realm of weird and keyword-packed attempts at creating recommendation-optimised videos, it's almost inevitable that you'll run into content that most parents would rather not expose very young children to.

By clicking on content from YouTube's suggested video bar, we went - via weird children's YouTube - from an entirely legitimate CBeebies video to a low-budget Paw Patrol rip-off showing one of the series' canine heroes attempting suicide by jumping off a building A random browsing test with a toddler who wasn't actually shown any unpleasant content took us from YouTube-native children's favourite Bob the Train to, a few stubby-fingered lunges at the tablet screen later, a fake Mickey Mouse cartoon with depictions of eye-gouging, a parasite-infested stomach and small children setting each other on fire.

High viewer counts are a common feature of this more distressing content, too. Finger Family Song Nursery Rhymes racked up 20,, views in two days before we reported it. Finger Family Song Nursery Rhymes — which includes a brawl and a depiction of one of the cartoon characters mourning over his friend's grave — was at over 11m.

It takes a few unfortunate turns to get from generic weird children's YouTube to genuinely distressing content, but anyone who's spent much time on the platform will notice that certain videos are recommended over and over again, unrelated to their apparent quality and frequently appearing — even if you express disinterest in them using YouTube's limited viewing management options.

The disproportionate likelihood of some videos to be recommended, whether to a user with no YouTube history or someone with established viewing habits, implies that there are factors unrelated to the behaviour of the logged-in user that influence why YouTube's suggestion algorithm favours certain content. The company is obviously motivated to keep details of how its algorithms work secret, not least of all to prevent them from being exploited.

However, YouTube global head of public policy Juniper Downs' statements — to the Parliamentary inquiry into fake news in February — provide some insight. There is also content that is associated, so if I have a real niche interest in a particular kind of knitting, there may not be a lot of highly popular videos that are about that type of knitting, but we will continue to recommend videos that are similar and that provide more instruction on that type of knitting.

YouTube's position is that gaming its recommendation system would be extremely difficult. However, the way certain channels are flooding the platform with huge numbers of largely interchangeable, low-value videos that purport to be educational would be regarded as spam at best in any other medium. The junk content problem becomes particularly visible in the restricted environment of the YouTube Kids app, where a chain of recommendations can rapidly devolve into endless computer-generated models of trucks, animals and superheroes, cartoon head-swaps and voice samples yelling colours.

It's unlikely to do any harm, but content of this kind is disconcerting and a long way from the enriching and educational environment that YouTube obviously wants to create for its youngest viewers. YouTube removed or restricted unsuitable videos reported during our tests, typically in a matter of hours. But case-by-case removal isn't a suitable long-term solution for dealing with huge amounts of content that's at best weird and at worst potentially disturbing for its very young target audience.

Since the publication of Bridle's November , Medium post drawing attention to the problem, YouTube has taken measures to improve its detection of genuinely harmful content targeted at the youngest section of its audience.

These include demonetising — removing ads — from "any content depicting family entertainment characters engaged in violent, offensive, or otherwise inappropriate behavior, even if done for comedic or satirical purposes" and blocking the most inappropriate videos.

YouTube spokesperson Thea O'Hear says that the service is now using a combination of machine learning and human expertise to flag up dubious videos: "YouTube is doing more and more enforcement automatically using machine learning, but we will always need people to flag [problematic content] to us as well. The company's Trusted Flagger scheme gives charities and government agencies access to tools that allow them to bulk flag problematic content, which YouTube staff will then review as a priority.

These include child safety experts, who also provide the company with guidance on new threats and trends as they evolve. Despite reams of dubious content aimed at children on YouTube, we don't know what impact it has.

On the other hand, several meta-analyses have found no effect at all — watching violent media did not lead to aggressive or violent behaviour. It's normal for parents to be concerned if they find that their children have been exposed to content that depicts violence, harm or other bad behaviour.

Children may also try to process what they've seen through play, for example by making their toys fight, or may use sexualised or violent language that worries the adults in their lives.

However, the way parents and carers respond to such situations is key and can make a big impact on how children understand what they've seen. While the UK doesn't have any official guidance on screen time for young children, he points to the American Academy of Pediatrics' guidance that: "For children aged 2 to 5 years, limit screen use to 1 hour per day of high-quality programs.