Teenagers are experiencing a mental-health crisis. And though the science is messy and the matter isn’t settled, many suspect that social media is, in some substantial way, tangled up in the problem. Following this instinct, legislators and regulators at both the state and federal levels have suggested a slew of interventions aimed at protecting young people from the potential harms of social platforms. Many of these efforts have so far fallen short on legal grounds, and broadly speaking, the status quo remains.

This week, we learned of a new approach intended to protect kids from Big Tech. On Tuesday, a joint lawsuit was filed against Meta by the attorneys general of 33 states, deploying consumer-protection laws to try to hold the company accountable for harming young people. It claims that Meta deliberately got children and teenagers “addicted” to its platforms, that this addiction directly causes physical and mental harm, and that the company lied about it.

[Read: Why American teens are so sad]

“Just like Big Tobacco and vaping companies have done in years past, Meta chose to maximize its profits at the expense of public health,” Colorado Attorney General Phil Weiser said in a press release. The Big Tobacco comparison has been made a number of times since fall 2021, when the whistleblower Frances Haugen leaked to the press internal Meta documents about Instagram and Facebook. Among them were the results of studies showing teenagers candidly reporting the negative effects that social media was having on their lives. When they felt bad about their bodies, Instagram made them feel worse. They had noticed increased anxiety and depression among their peers, and they considered Instagram to be one of the causes. Haugen’s so-called Facebook Files preceded the attorneys general investigation, which were announced several weeks after their release.

The suit is worth reading closely. As an effort to address incredibly serious social problems, it’s surprisingly slapdash. Our window into the case may be limited—many of its 233 pages are at least partially redacted, some blotted out entirely—but what is visible clearly relies on familiar, flawed tropes. It doesn’t engage seriously with the thorny question of just how social media affects kids and teenagers, and instead reads somewhat like a publicity stunt. Experts told me that the legal arguments made in the suit, even without knowing what is in the redactions, are not particularly convincing.

“I’m sympathetic overall to the dangers that social media pose to kids and how platforms have been poor stewards of their responsibility,” Mark Bartholomew, a professor at the University at Buffalo School of Law specializing in technology and the law, told me. “But when I look at the law … I do think it’s a stretch.” There are a couple of problems, he said. First, although social-media use might be a compulsive behavior, there is no official diagnosis for such a thing as social-media addiction. Second, proving that deception played a role in consumers’ use of Meta products will also be a challenge. That argument hinges on Meta’s public assurances that its products are safe, as well as the notion that consumers have taken that at face value to the point where they have been genuinely misled. “It’s hard to show that people were deceived,” Bartholomew said. “That they thought Instagram was one thing and it turned out to be another.”  

In connection with these arguments, the suit puts forward the idea that Meta deliberately presents young users with content that will “provoke intense reactions,” such as “bullying content” and content related to eating disorders or violence. The problem with these arguments isn’t that they are unfair; it’s that the notion that Meta would deliberately hurt the people it wants to keep on its platforms is both extremely hard to prove and easy to deny. (Young people absolutely are bullied through Instagram, and they certainly might see harmful content there—as with any internet platform, it’s impossible to argue otherwise. But does Meta display such material on purpose to lock users into the platform? Not exactly.) “Teens don’t want to be exposed to harmful content or hurtful interactions, and advertisers don’t want their ads showing up alongside content that isn’t appropriate for teens,” Liza Crenshaw, a Meta spokesperson, told me, arguing that the attorneys general had misunderstood Meta’s “long-term commercial interests.”

[Read: No one knows exactly what social media is doing to teens]

Experts agreed that another aspect of the case feels considerably more cogent: namely, that Meta has violated the federal Children’s Online Privacy Protection Act. “That part’s more concrete,” Bartholomew said. “At least, it’s a little harder for Meta to wriggle out of.” COPPA prohibits tracking the online activity of children under the age of 13, or collecting their personal information, without explicit parental consent. If Meta has what COPPA terms “actual knowledge” of kids younger than 13 using its services, it’s violating the law. (“Instagram’s Terms of Use prohibit users under the age of 13. When we learn someone potentially under 13 has created an account, we work to remove them if they can’t demonstrate they meet our minimum age requirement,” Crenshaw said in a comment.)

Berin Szóka, a lawyer and the president of the libertarian-leaning think tank TechFreedom, highlighted one place where the suit’s argument could hold water: the complaint that, on Instagram’s sign-up page, where it asks for a new user’s birthday, the menu previously would automatically suggest a birth date 13 years prior. “That’s not a neutral age gate. That encourages the answer of Yes, I’m exactly 13 years old,” he told me. Meta recently changed this age gate, but it could be fined retroactively, and the attorneys general could ask to have some kind of continued supervision of the company’s COPPA practices. This would be a significant win, even if other elements of the suit are dismissed.

Most of the details in this part of the suit are redacted, so it’s possible that the states found new evidence of current lawbreaking activity as well. What is visible to the public so far is a bit ridiculous, however. For instance, to prove that Meta knows that kids use its apps, this suit cites the simple fact that various kid-oriented brands and media personalities (Lego, Hot Wheels, SpongeBob SquarePants, JoJo Siwa) have Instagram pages. The evidence in a similar (settled) case against YouTube was far more direct: While publicly denying that kids used YouTube, YouTube was also taking meetings with toy companies such as Mattel and Hasbro and literally pitching itself as a “leader in reaching children age 6–11,” as well as the “#1 website regularly visited by kids.”

Where does this leave us? Mostly, wondering what broader outcome the states are hoping for. The attorneys general say Meta has used “powerful and unprecedented technologies” to “ensnare” youth and teens. That might be a common rhetorical point in popular discourse, but it would require a lot of work to prove. And by far the weakest part of their argument comes when the states try to substantiate the claim that, as New York Attorney General Letitia James said in a press release, Meta is “to blame” for the mental-health crisis among kids and teenagers.

In the clearest statement of their position, the attorneys general write: “Increased use of social media platforms, including those operated by Meta, result in physical and mental health harms particularly for young users, who experience higher rates of major depressive episodes, anxiety, sleep disturbances, suicide, and other mental health concerns.” There is just one citation on this line, to a public Google Document maintained by Jonathan Haidt, a social psychologist at the NYU Stern School of Business (and a contributor to The Atlantic). That document summarizes dozens of studies with different findings, some of which contradict one another. Which ones are the attorneys referring to? They don’t say.

Later, they do cite a specific 2022 study by Amy Orben and Andrew Przybylski, well-known researchers in the field. They found that young people are vulnerable to decreases in life satisfaction (quantified with a questionnaire) as a result of excessive social-media use in particular age windows. (For girls, ages 11 to 13; for boys, ages 14, 15, and 19.) In the lawsuit, the attorneys summarize the study as finding that “going through puberty while being a heavy social media user interferes with a sensitive period for social learning.” This is not an accurate representation of that study at all. “We did not show that social media interferes with social learning,” Orben said when I emailed her the page of the lawsuit that cited her paper. In fact, the words social learning don’t appear in the study at all.

Bartholomew offered a theory of the case. “AGs get a certain amount of deference in the courts,” he told me. This isn’t a private class-action suit that can be quickly thrown out. “It’s unlikely to be dismissed anytime soon, and I think the main point here is to make some waves.” Maybe that’s fine. But neither the mental-health crisis nor the expansive power of social-media companies will be seriously dealt with this way. Whatever the intentions of this suit, it’s not striking anywhere close to the crux of our problems.

This is an edition of The Atlantic Daily, a newsletter that guides you through the biggest stories of the day, helps you discover new ideas, and recommends the best in culture. Sign up for it here.

Terrorism, like war, is a word we tend to use almost as a reflex to describe anything that horrifies us. But words can lead us to choose policies, and we should be aware of how we use them.

First, here are three new stories from The Atlantic:

“Why I own guns and why I’m part of the problem.” Hurricane Otis was too fast for the forecasters. Will Republicans pay a price for extremism? The hero Gen Z needs

Another Terrifying Day

As I write this, a mass shooter is loose in Maine. I have close family members who live not far from the scene of the massacre, and, like all Americans, I am praying that his rampage is stopped before he kills again.

I do not know why someone in Maine engaged in a mass slaughter yesterday. (Authorities have identified a suspect, but I see no point in naming him here.) The alleged shooter was reportedly committed to a mental-health facility this past summer, but I do not know what condition led to his stay. I do not know if there was some precipitating event, or whether he was under the influence of drugs, or if he is just an evil human being.

I also do not know if he is a terrorist. At this moment, no one does. But on social media, especially, the word terrorist is being thrown about with great confidence, especially now that we have some evidence that the suspect’s social-media feed was heavy with likes of right-wing accounts. This may not mean much; the alleged shooter also seemed to like Jim Cramer and other finance-related accounts. We can’t really ascribe motive out of any of that; sometimes, people are radicalized and become dangerous, but other times, dangerous people seek out causes as a rationalization for violence.

I will be honest here and tell you that I considered leaving this subject for another day. We’re all scared, shocked, and angry. But times like this, when our fears are so sharp, are exactly when we need to think more calmly about the nature of the threat we’re facing. When we rush to apply words because they seem right to us in the heat of the moment, we run the risk of making mistakes that will reverberate throughout our later discussions and influence the policy choices we eventually make.

The U.S. government has its own definition of terrorism, and it is fairly loose—not least because after 9/11, the government wanted more flexibility in charging people for terroristic acts. But let’s start with something very important that almost all governments agree on: Terrorism is a political act intentionally aimed at civilians in order to produce fear and subsequent changes in government policy (or even the destruction of the targeted regime).

Usually, definitions of terrorism emphasize that the perpetrators are nongovernmental actors, because we already have terms for when states engage in the intentional murder of civilians: crimes against humanity and, in some cases, war crimes. (Intention is important: Civilians are always killed in wartime, but specifically targeting them is a crime.)

Counterterrorism operations also look for networks, planning, and cooperation among the killers. These networks have goals: Sometimes, the goal is relatively achievable (“release our comrades from prison”), sometimes it is huge (“give us autonomy” or “remove your forces from this area”), and sometimes it is nearly impossible (“overthrow your government and adopt our religion”). But there is always a goal.

Terrorism without a political motive isn’t terrorism. Not everything that terrifies people is terrorism, either, as counterintuitive as that may seem. After all, if it’s terrifying, it’s terrorism, right? Nevertheless, although many things scare (and kill) large numbers of people—gang wars, serial killers, arson—those that lack a coherent political character fall outside the legal, and sensible, definition of terrorism. They are crimes against other human beings, but they are not an attack on the entire political order.

Why does any of this matter? Above all, we need clarity on the nature of the crime so that we can choose the right response. Ever since 9/11, invoking terrorism in America has carried the possibility of setting in motion the immense machinery of government, regardless of the actual threat. But if we more carefully define terrorism to mean non-state actors attacking civilians to produce a political outcome, it gets a lot easier to think about how to react.

For example, Son of Sam killing six people, wounding seven others, and scaring the hell out of New York in 1976 and 1977 is ghastly, but it is not terrorism. But a car bomb in front of a mall—or a jetliner aimed at a building—attached to a political or social cause is terrorism. Son of Sam requires a manhunt by local and regional law enforcement. The car bomb requires a significant governmental response—and perhaps even military mobilization.

The shooting in Maine is not the only event spurring the daily deployment of terrorism as a term. The Hamas attack on Israel is now “Israel’s 9/11,” and the United States is reportedly advising the Israeli government not to make some of the same mistakes America made in its own War on Terror. (War is another term thrown about too easily, but that’s a subject for another day.) I know the old saw “one man’s freedom fighter is another man’s terrorist,” but I believe that the Hamas invasion was terrorism: Non-state actors intentionally targeted civilians to effect a political goal.

You can argue over justice and morality—some people have made the despicable argument that Israel brought this nightmare on itself, similar to arguments made about America deserving what happened on 9/11—but there can be no argument that rape, infanticide, and butchery in service of a political goal are terrorism. (Russia has done the same in Ukraine—but as a state actor, the Kremlin and its high command should be charged with crimes against humanity and war crimes.)

In Maine, the situation is far less clear. It might make us feel better, and give more meaning to the heartbreaking deaths, to believe that we’re fighting terrorism; the alternative is to wrestle with the even more frightening and desolating possibility that the Maine shooter may (like the Las Vegas killer in 2017) have had no real reason to kill beyond his own unknowable inner torment.

When we use a word such as terrorism promiscuously, we risk turning it into little more than shorthand for our fear and anger. The term not only invites a massive government reaction but could also lead to misallocation of resources in our responses, especially if we conflate mental illness, the obvious problem of guns, and “terrorism.”

To take but one example: In late 2021, a mentally disturbed 15-year-old named Ethan Crumbley killed four people at his school. He was convicted of murder—and of terrorism, under a state law enacted after 9/11. (The prosecutor’s argument was essentially that Crumbley’s act had terrified people, and so: terrorism.) If a teenage school shooter who was hallucinating about demons and sending messages pleading for help is a terrorist, then the word has virtually no meaning.

Sanctifying the word terrorism as an obvious motive for every mass killing was a significant mistake made by Americans and their government after 9/11. The world is crawling with plenty of real terrorists, but we should pause before we reach for a word whose incantation can summon powerful and illiberal forces from within our institutions—and ourselves.

Related:

The narcissism of the angry young men A lone-wolf shooter has an online pack.

Today’s News

Israel sent armored tanks into northern Gaza overnight following remarks from Prime Minister Benjamin Netanyahu about a likely ground invasion.    Representative Jamaal Bowman pleaded guilty to setting off a false fire alarm in a House office building. The Texas House of Representatives passed a bill that would make it a state crime to cross illegally into Texas, and enable officers to arrest and deport undocumented immigrants.

More From The Atlantic

Dobbs’s confounding effect on abortion rates Biden says goodbye to tweezer economics. “I love candy. But does it make me happy?” We’ve never seen anything like the Menendez indictment.

Culture Break

Read. They Called Us Exceptional: And Other Lies That Raised Us, a memoir by Prachi Gupta, delves into the grief of cutting off family, and argues that estrangement can be a tool of self-love.

Listen. In the latest episode of Radio Atlantic, host Hanna Rosin speaks with Jordan Peele and N. K. Jemisin about their new anthology, Out There Screaming, and the subversive goals of Black horror.

Play our daily crossword.

P.S.

I am a traditionalist who dislikes much about modern music. (I think Auto-Tune is a crime against God and man.) So I cringed when I saw in The Guardian that Paul McCartney and Ringo Starr are going to use AI to resurrect John Lennon for one more Beatles tune, with a guitar part recorded in the 1990s by George Harrison, who died in 2001.

When I learned more, I was heartened. I liked the use of John’s voice in later Beatle releases, especially the song “Free As a Bird.” Apparently, John’s widow, Yoko Ono, had some of these materials on a cassette John had marked “For Paul,” and the three surviving Beatles at the time used modern studio magic to clean up the tapes. But technological limitations prevented them from using all of John’s singing and playing. AI allowed Paul and Ringo to restore his parts in the new single, titled “Now and Then.”

George reportedly didn’t like “Now and Then,” but his widow and his son think that with the restored quality, he’d have approved. It wouldn’t be the first time the Beatles disagreed on a song. But I’m glad we’re going to get one more single from them before they finally close their legendary catalog.

— Tom

Katherine Hu contributed to this newsletter.

When you buy a book using a link in this newsletter, we receive a commission. Thank you for supporting The Atlantic.