A few weeks ago, I happened to be watching a football game on TV. Between the bouts of gridiron action were commercial breaks. This was, obviously, no surprise. Television advertising has been around since 1941 and has since grown into a billion-dollar industry.
And yet, sandwiched between the other advertisements hoping to capitalize on the attentions of football fanatics everywhere, one particular commercial caught my eye. This ad for Coca-Cola:
A Christmas commercial from Coca-Cola. Red trucks, Christmas trees and polar bears. Pretty standard stuff from the soda giant. So, what’s the big deal?
Well, I couldn’t shake the feeling that something seemed just a little off about this commercial. To be honest, I thought it looked like it had been made with AI.
Turns out, it was.
To be fair, Coca-Cola wasn’t totally trying to hide that fact: There’s a tiny disclaimer that pops up at the start of the video, acknowledging that the ad was made by artificial intelligence (“Real Magic AI,” in this case). But that notice isn’t prominently featured, either.
Others picked up on this artificially created commercial, too, and I think it’s fair to say that the public response probably hasn’t been what Coca-Cola had hoped for. Still, it serves as a glimpse into the power AI video generators wield and how even major corporations—businesses that certainly have the resources to hire actual flesh-and-blood creators—intend to use them.
But these AI tools aren’t only available to big companies. There are plenty of free AI video generators that can be found online—and they’re pretty popular. According to Semrush, “ai video generator” is searched for online nearly 75,000 times a month! And that’s to say nothing of deepfakes, which is its own branch of AI-generated video.
For the purposes of this article, we’ll stick to AI generators that “create” videos seemingly from scratch, like that Coca-Cola commercial. Because these tools are easily accessible online, both for adults and kids.
How Do AI Video Generators Work?
Videos created by artificial intelligence falls under the banner of “generative AI.” According to Harvard University, generative AI is “a type of artificial intelligence that can learn from and mimic large amounts of data to create content such as text, images, music, videos, code and more, based on inputs or prompts.”
A few weeks ago, we talked about how generative AI works when creating images. AI videos work in a similar way. That is to say, the AI video models use a huge deposit of reference data to form their “creations.” But there is a major hurdle that the model must account for when doing video instead of images—movement.
Generative AI & Video
To create movement, the AI model looks at the connection between multiple images in a sequence and how the subjects in those images transform. For instance, imagine multiple snapshots of a person running on a track. Each picture would show the person in a different position as they ran. Now, you or I would have a general understanding of how that person moved from one position to the next. That’s because we understand the function of running—what it should look like and what it shouldn’t.
But an AI model doesn’t have that understanding. It can’t run for itself. It can’t even use its eyes to watch somebody run because it doesn’t have eyes to see. The best it can do is process the given images—analyze the difference between individual pixels—and try to decipher how a running person moved from one position to another. It follows the same process when trying to replicate any movement in generated videos.
Sometimes, the AI will get close to accurate. In the aforementioned Coca-Cola ad, the generated video could easily pass as a schmaltzy Christmas commercial to those either unfamiliar with AI or not paying close attention. The ad cuts quickly between scenes so as not to tax the limits of believability, but its peculiarities are not glaringly obvious—at least, not at first.
Other times, though, a generated video might seem downright inhuman—especially when featuring, well, humans. Characters morph into different versions of themselves from scene to scene (as exemplified by the following AI-generated Toys “R” Us ad from earlier this year). Objects don’t function as they should—such as a car rolling down the road despite its wheels remaining perfectly stationary. And, at times, it can also seem that the subjects of AI videos are unconstrained by the laws of physics.
Kids & AI Video Generators
It’s important to be able to distinguish between what’s real and what’s generated by AI because AI video generators are not going away anytime soon. Likewise, it’s important for parents to know if their children are using these tools.
According to Common Sense Media, approximately 70% of teens have used generative AI tools, and most teens have done so without their parents’ knowledge.
Potential Benefits
Similar to AI-generated images, children can use AI tools to create videos in seconds. They can play around with many different scenes, characters and styles in short videos, which could spur their imagination and lead to other creative ideas. Generative AI could allow them to create and share something which only existed in their minds before, which can be an encouraging experience.
And, depending on the user, the dreamlike quality of some AI-generated videos is a feature, not a bug. If your child delights in more fantastical images and stories, AI videos could encourage them to expand their imagination in creative and interesting ways.
Potential Concerns
Unfortunately, AI video generators share many of the same concerns as their algorithmic cousins. Research is still being conducted on how the use of AI can affect a child’s creative process—it would be awful for a creative tool like generative AI to negatively impact kids’ creativity.
Additionally, AI is no replacement for experience. After all, that’s what AI struggles with, right? It makes mistakes because it can’t understand function: It has never physically experienced what a human being has. So, you should encourage your kids to experience things that they can then draw from when creating their own videos.
Finally, as with other generative AI tools, bad actors may misuse video generators to create disturbing or explicit content. Most tools have some sort of safeguards in place, but those are far from foolproof. A child who is interested in AI video generators could be exposed to harmful videos if they take a wrong turn.
What Can Parents Do?
So, as a parent, how should you approach this topic with your kids? First, make sure the lines of communication between you and your child are in good shape. You’ll want to ask them if they’re aware of AI video generators. If they seem interested, offer to explore some of the AI tools together.
If they happen to have already used generative AI for videos, ask them to show you what they’ve done with it. Foster an open and honest dialogue with your child and try to help them understand when an AI video generator might be appropriate to use and when it wouldn’t.
At the end of the day, if you choose to allow your kids to use video generators, be sure to encourage their own individual creativity apart from AI. Every child is unique, so decide for yourself what’s best for your family when it comes to generative AI.
