Jonathan Haidt on Social Media, Kids, and the Fight Over Who Pays for Harm

Jonathan Haidt has a way of describing what happens when a small child is handed a touchscreen: the child touches something, gets something, and quickly learns to touch in order to keep getting. In a recent New Yorker interview with David Remnick, the social psychologist and author of "The Anxious Generation" returned to that idea as a trial opened in California against social-media companies, and as lawmakers elsewhere, Australia among them, moved to shield children from the platforms.

Haidt's book, published two years ago, became a best-seller and has kept its influence. It argues that social media has delivered a mind-altering danger into childhood. Remnick notes that Haidt was at first met with critical eye-rolling, partly from researchers who thought some of his claims were not fully proven, and partly from people who saw him as an alarmist complaining about kids and their devices. As evidence of social disconnection and declining youth mental health accumulated, the book became essential reading for many.

The California case, Haidt explained, runs into a law from the nineties. Section 230 of the Communications Decency Act gave internet companies near-blanket immunity from being sued over what other people post. Congress meant it to free companies to take down material such as pornography without becoming liable for everything they left up. Courts, in Haidt's telling, have read it far too broadly. He described parents whose children died or were harmed, including a child who was sextorted on Snapchat and died that night, and an eleven-year-old girl who joined Instagram happy and developed an eating disorder within weeks. By his account, none of those parents has had a day in court against Meta, and none of the companies has ever faced a jury.

He offered figures. Snap itself acknowledged receiving ten thousand sextortion reports a month from users in 2022, and, in the company's own words as Haidt relayed them, that was likely the tip of the iceberg, since many victims do not report. He expects artificial intelligence to push the numbers higher. On mental health, he said that when surveyed, twenty to thirty per cent of girls report that social media harmed theirs, a scale he considers plausible as a cause of the rise in mental illness beginning around 2012.

The lawsuits cannot be merged into a class action, Haidt said, because class actions require plaintiffs to have been harmed in the same way, and these stories differ. Instead, thousands of cases have been gathered as multi-district litigation before a single judge in California. Because hearing them all is impossible, both sides argue over which cases to try first. Those bellwether cases go before juries, and their outcomes should signal how the rest are likely to go. The result Haidt hopes for is a jury finding that the platforms are addictive and were designed to maximize engagement. As an example of the design tricks, he pointed to the pull-down-to-refresh gesture on the iPhone, which he said was copied directly from slot machines.

Remnick raised the objection he hears from younger colleagues: that this is one more moral panic, like those over comic books, television, or rock and roll. Haidt called that a legitimate argument he has to answer. His answer is behaviorist conditioning. Television has no stimulus-response loop, he said; a smartphone is closer to a Skinner box, in which B. F. Skinner used variable-ratio rewards to control the behavior of pigeons and rats. He also cited surveys in which young people say the platforms harm them, and noted that half say they wish TikTok had never been invented. Nobody, he said, made that claim about comic books.

Remnick also described meeting Apple leaders, who he said were blasé and essentially told him to turn the device off and ration his time. Haidt responded that when everyone behaves in a way that seems bad, the situation is probably at fault. Addictive products like gambling, alcohol, and cigarettes are, he said, kept from children, even if adults are trusted to regulate themselves despite the roughly ten percent who may be severely damaged.

The conversation then turned to reading. Haidt now believes attention damage may be the biggest harm. He contrasted "slow dopamine," the satisfaction of a child slowly finishing a tree house after failures, with the quick dopamine of swiping away from anything uninteresting within eight seconds. He described this as the subversion of the ability to pay attention on a species-wide level. Remnick, a professional reader, said he puts his phone in the kitchen to read manuscripts in another room. The bellwether verdicts in California are the next thing to watch.