Calm Down About the Kids

“Speaking as a parent,” I end the year with some words of reason about teens and technology.

I recently spoke on a panel about kids, social media, and artificial intelligence. We were forty minutes in before I said a word. There was, to be fair, a keynote from a member of Congress, which took up a good chunk of time. Then a researcher spoke. Then an activist. By the time it got to me, we had cycled through every social-media and AI panic talking point: addiction, mental health, sextortion, bullying, distraction. I’m sure I’ve missed a few.

My experience is illustrative of a broader trend. Consider a recent hearing before the House Energy and Commerce Committee, which convened to discuss a slate of nineteen bills dealing with minors and the internet. In the words of one representative:

Facebook is the Philip Morris of our time. Yet they are just one example of the cesspool that is the modern internet. Americans are waking up to Big Tech’s game. They provide lip service on how they’re protecting children, while actually delivering none of those protections. … Countless studies and multiple congressional hearings make clear that social media is not only [an] addictive service for children, but is also harmful for their mental acuity and overall mental health. The damage social media is doing to our kids is unconscionable and extremely disturbing. ... As a father of three school-age boys, the warning signs of social media’s impact on kids are abundantly clear to me and to all parents.

Yeah, so that’s crazy talk. But it was the tone of the entire hearing. A lot of grandstanding, a lot of demagoguery, a flight from nuance of any kind. The adults were the children in the room, insisting that you take their dramatic outbursts seriously.

I’ll admit to feeling a little gloomy about tech policy as we head into the new year. It seems like the tide is shifting. It seems like we’re about to see a more closed internet.

But resignation gets you nowhere in this world. So let me end the year with some words of reason.

If you are not a parent, do not let elected officials waving their children around intimidate you. A popular tactic is to say, “As a parent with kids … ,” and then to act like whatever words follow carry great moral authority. But many of these bills would lock down the internet; they would place age-verification barriers between you and your speech rights. You may not have children, but it is your rights on the line. Don’t be afraid to speak up.

If you are a kid, your speech rights are at stake too! Be not afraid. Don’t let all the panic scare you. Beware an expectancy effect, where you wind up feeling anxious or guilty about social media or AI simply because the prigs, the new Victorians, tell you you’re supposed to feel that way.

And if you’re a parent, keep calm and carry on. There’s a lot of irresponsible talk flying around. Most of it is nonsense.

Is social media addictive for your kids in the way nicotine or heroin is addictive? No. A recent Pew survey revealed a modest decline in young people’s use of platforms like YouTube and TikTok. Kids are spending more time on chat apps—WhatsApp, Discord—talking to one another. This should not be surprising. Teens are social creatures. They want to talk to their friends. They are apparently not uncontrollably hooked on these products. They are apparently perfectly capable of using them to serve their own ends.

That lines up with the research. Strip away the heated rhetoric, and the debate boils down to whether social media has no net effect or a small net negative effect. The outsized alarm reflects the fact that we are in the middle of a moral panic. We are paying enormous attention to social media. If we devoted the same level of attention to the Boy Scouts or chess clubs, we would probably find similar results. Many kids benefit; many are not much affected one way or the other; a few have bad experiences. A handful suffer some kind of tragedy—say, abuse at the hands of an adult.

Kids are a diverse lot. We need to account for individual experiences. Many young people benefit greatly from online communities: they escape hostile offline environments, find like-minded peers, and build forms of solidarity and support. At panic-driven hearings, those kids are invisible.

The negative effects that do exist should not be ignored. They are real, even if overstated. But they are wildly outsized in the conversation right now.

Personally, living in California, I would much rather focus on the education system. This is a state that blithely embraces whole-language reading, equity-based math, discovery learning, and other harebrained experiments in pedagogical regression. The educational establishment maligned phonics, ditched standardized tests, and inflated grades. San Francisco tried to abolish eighth-grade algebra and tracked learning. There’s a pervasive culture of failure. (The people who holler loudest about social justice are often the most energetic backers of policies that inflict immense harm on our most disadvantaged children.)

Our kids have regressed decades in math, and now graduate high school several grade levels behind kids in Japan or South Korea. And you want to talk to me about screens? They have screens in East Asia too, by the way.

“Speaking as a parent,” I’d rank the existence of social media, on my list of concerns, about where I’d rank whether the schools have soda machines. Yes, some kids will love soda and harm their health binging on empty calories. If we treat every temptation as an addiction, there is no principled stopping point to what we’ll have to ban.

Again, this is not to say that social media is never harmful. But we need to undertake a sober accounting of relative risks. The narrative has gotten way, way ahead of the evidence.

The same dynamic is now playing out with AI. That’s how this works. A new technology comes along. The kids like it. The adults freak out. Comic books, television, video games, social media, AI. It happens over and over. Joe Lieberman holds up the blue toy gun at a Senate hearing, claiming that video games turn kids into violent psychopaths. And before the evidence is properly collected, considered, and placed in context, we move on to the next thing.

Part of the problem is that you can tell a mechanistic story about why these things are harmful, and the story will hang together. At the panel I was on, the researcher talked about children’s brain structures and how their development might be affected by talking to a chatbot. It sounded kind of persuasive, in the same way that I could tell you a persuasive-sounding story about how video games affect the developing brains of children in a way that will make them violent.

But if you want to know whether a technology is actually causing harm, you have to look at cohort studies and tease out whether there is a causal relationship having a concrete effect. And we generally don’t do that. We are awash in studies that speak in the language of correlation. Serious researchers like Candace Odgers show that, in all probability, it’s depressed children who use social media more, rather than social media use that makes kids depressed. But those findings get drowned out in the shouting. (Tobacco! Philip Morris! They’re doing nothing!)

Yes, there have been tragic cases involving chatbots and vulnerable users. We should absolutely discuss the state of AI’s guardrails. The companies themselves largely agree. But the idea that chatbots can cause suicide is simply unsupported. All we really know is that if a child is suicidal, a chatbot cannot be relied on to stop him from committing suicide. That shouldn’t come as a surprise.

As a parent, you should talk to your kids about these tools. Supervise them. Pay attention. But as with social media, we’re papering over the wide array of different experiences that young people are going to have with the technology.

There is now a push to cut children off from AI outputs entirely. That would be a mistake. LLMs, including chatbots, can be genuinely useful. Kids can ask questions they wouldn’t ask an adult. They can explore concepts without fear of shame or stigma. They can use these discussions to mature in their self-conception. They might even (heaven forfend!) have chatbot friends. As long as those do not drastically displace real-world friendships, there’s no cause for alarm.

That’s another thing about all this addiction talk. It robs young people of agency. It makes them sound like they’re brainwashed or manipulated by any technology they interact with—as though they can’t make their own choices or form their own preferences.

There is also a strange double standard at work. Technology is treated as suspect unless it’s perfect. The chatbot might have biases. Well, humans have biases! Frankly, I worry sometimes about the biases that adults bring to conversations with my children. Chatbots, by contrast, are often remarkably good at presenting information in a neutral manner. My children and I talk to chatbots together. We learn a lot.

Here again, the evidence does not justify the hysteria. A recent OpenAI-MIT study found that roughly 0.15 percent of users show signs of unusually strong emotional attachment. That is worth paying attention to. But again, a lot depends on what we choose to look at. I’m confident that we could take any number of activities in society and find that 0.15 percent—the power users of almost anything—have a problematic relationship with it. People form weird obsessions about things. That tells us more about people than about the mind altering power of the things they interact with.

If I could propose a New Year’s resolution for Congress and the activist class, it would be this: Have some perspective. Stop panicking for political points.

Most kids are doing okay. And to the extent they’re not—it’s true that the mental health figures for teens aren’t great right now—I’m sorry, but you’ll have to reach for explanations beyond technology. You might have to grapple with the divorce rate. Jonathan Haidt—who I guess contains multitudes here—is worried about helicopter parenting. (He also seems to think that maybe you should try not being liberal?) Abigail Shrier contends that this is all iatrogenic, and that we need to stop encouraging kids to pathologize themselves. Derek Thompson wonders if there’s something in the water in Anglophone countries, since many non-English-speaking developed countries, complete with smartphones, have not seen the same mental-health decline as we have.

I’m not sure what to make of any of these explanations. What I do know, however, is that it’s an uphill climb to show that technology is an independent cause of mental health problems.

If you’re a parent, tell your kids you love them. Tell them they can talk to you. You are probably doing better than you think. It’s worth remembering that, the bad mental-health trends notwithstanding, people are, by and large, who they are. The blank slate theory is wrong. Your kid won’t be derailed by your every little mistake. If your child uses social media or AI, chances are they’ll be just fine. The genes will out—yet another simple truth that gets lost in a moral panic.

Expose your kids to technology gradually. Let them make mistakes, just as they do in the physical world. Let them learn. Let them grow.

Happy New Year.