The internet needs crash test dummies
More and more governments are banning social media for teenagers, despite mixed evidence for their harms. Perhaps a better idea would be looking at car safety.
Apologies for newsletter silence recently. I’d love to say I’ve been busy, but the reality is that I’ve been stuck in a bit of a creative dry spell. (Blame the heatwave.) Forgive me if I’m rusty.
Finally, a solution to all the internet’s problems: ban men. After all, adult males are the biggest perpetrators by far of online abuse, grooming, child pornography, and creating and spreading misinformation. Don’t ban teenagers, ban men!
When I read this idea, suggested in an interview with the psychologist Candice Odgers in the Guardian, I laughed. Odgers’ solution is a funny (in a joking-not-joking) way to cut through the hyperbole surrounding social media regulation, which has now reached castrato pitch. The British government’s latest masterstroke is announcing that it will impose an internet curfew for Under-17s1, an extension of the ban on social media for Under-16s passed last month. It’ll surely ban dancing next.
The Brits aren’t the only ones. The European Commission has this week pledged to introduce an EU-wide social media ban for children, following up the likes of Australia, Brazil, France, Turkey, and China, all of whom have announced or enacted some form of social media ban for teens. Blame Jonathan Haidt, but governments have at least finally coalesced around an approach to regulating the internet: ban now, work out how later.
It’s worth pausing to note what a marked shift this is. Western governments, and the British government in particular, spent the 2010s in thrall to nudge theory: the idea that by using small tweaks, such as displaying calories on menus or reducing the size of serving spoons, you can influence the public make healthier decisions without needing to pass anything so draconian as bans. Nudge theory became the defining regulatory theory of the TED Talk era, a decade characterised by Silicon Valley idealism2 and charlatans building careers on oversimplifying entire fields into neat 15-minute slide decks.3
It wasn’t surprising to find that much of nudge theory — like “power poses” and other pop behavioural science — turned out to be mostly bunkum. Thus prohibition, the favoured tool of authoritarian governments, is back in vogue. And social media, being that rare bipartisan cause in these hyper-partisan times, is the number one target.
Look, I’m not pro giving teenagers social media. I have two kids of primary school age, and we limit their screen time as much as possible, and steer what usage they are allowed towards educational programming however we can. I fully believe that Big Tech intentionally designed social media platforms to be addictive and inflammatory; and that some have knowingly spread hateful and criminal content and imposed horrifying conditions on thousands of ill-paid workers in the name of content “moderation”.
But Odgers makes a compelling case that the current moral panic is not serving anyone. For one thing, the evidence of the negative effects of social media on teenagers’ mental health is mixed. And bans are, as policy, an extremely mixed bag. Even beyond the obvious failures (prohibition, abstinence, the war on drugs) there is little guarantee that they will have the desired effect. China, for example, tried to restrict kids playing video games too much, but its state curfew had little effect on the heaviest gamers. South Korea passed a similar law, but reversed the ban in 2022 after also finding it had little to no effect. In the UK, successive governments have passed increasingly draconian regulations targeting junk foods — banning advertising to kids, supermarket discounts, implementing a sugar tax, etc — in the name of making children healthier. But while the sugar content of fizzy drinks has fallen, overall obesity rates are still rising.
At the heart of this debate is an important question. What kind of ill is social media? Is it like tobacco smoking, which directly harms both the individual and those around them, via second-hand smoke? Or is it more like junk food — directly damaging to the individual, and indirectly to everyone else, via the toll on the health system? Exactly how harmful is it? Does it kill people? How many? Effective regulation hinges on these questions. And for social media, we just don’t have the answers yet.
I’m not a policy person, but I might suggest an useful model to consider: cars. For most of the 20th Century, cars were death traps. Car companies — giant, politically-connected behemoths that governments were loathe to regulate, lest they harm economic growth — fought every attempt at voluntary regulation: speed limits, seat belts, pollution controls. Then, in the late ‘90s, the UK’s Transport Research Laboratory launched the New Car Assessment Program (NCAP, which later spun out as EURO NCAP), a voluntary body that tested cars for crash safety. NCAP awarded individual car models between one to five stars, based on best-in-class crash testing. A bad NCAP rating sunk the worst-performing cars at launch, and in some cases forced manufacturers to change their designs, while helping to drive voluntary adoption for safety features like advanced braking and sensors. In the decades since, NCAP’s findings have saved countless lives.4
Importantly, NCAP did not pass regulations. It rewarded the best performers for producing innovative safety features, and penalised those that didn’t take safety seriously. Customers still had a choice — and they overwhelmingly chose models and companies that tested well, even if they were more expensive. (With cars, as with social media, people want to protect their kids.)
What would an NCAP for big tech look like? Perhaps a publicly-funded research organisation that tests social media platforms for harmful features — addictive algorithm designs, poor content moderation, illegal content — and gives them an official rating that users can use to identify what their risks are. Grok: may contain child porn. Instagram: may cause social anxiety. It wouldn’t prohibit users from using those platforms (though governments could still do so) but at the very least, we’d have evidence to build policy on. It could then compel companies to share features, like seatbelts, that are shown to benefit users the most.
This imaginary NCAP-for-tech would only become more helpful with the arrival of AI. Given that governments are locked in competition over AI supremacy (and already getting exclusive access to unreleased models, for “safety reasons”) an independent body would surely be preferable — or at least helpful for everyday consumers looking to make the right choices. In this age of disinformation, and agendas, and moral panic, wouldn’t it be great to feel a little more informed?
Besides, if we’re going to ban social media for kids, we should probably also do the same for their parents. A particularly compelling new study found that in Uruguay, the nationwide rollout of high-speed internet correlated with a marked decline in children’s development and social skills. The authors found that the primary cause wasn’t kids spending time on their phones — it was their parents, who spent more time on their phone and less time parenting. “When adults engage more with the internet, the quality and quantity of their interactions with young children is likely to decline,” the authors wrote. “A risk that should be emphasized as much as children's own screen exposure.”
Anyway. Time for me to put my screen away — I have to pick my kids up.
It is, hilariously, a “voluntary” curfew, teenagers famously enjoying both curfews and self-moderation.
In hindsight, almost all of the post-2008 Silicon Valley idealism was just a particularly aggressive form of Chicago School light-touch economics smuggled into the mainstream by men in hoodies and Patagonia vests.
A side note: the TED-talk-to-corporate-speaker pipeline has now been democratised by podcasts, which perhaps explains why similar events no longer hold much cultural weight, and are now largely private affairs for foreign investors.
Ironic postscript: while road fatalities have plummeted in recent decades, scientists have argued that the biggest driver behind the fall in road deaths is not seatbelts, or sensors, nor automatic brakes. It’s that parents don’t let their kids play in the street anymore, so instead of being hit by cars they’re inside… on screens.