For years, much of public health’s response to the well-documented mental health harm caused by social media among kids has been to tell parents to watch their kids more closely and tell them to spend less time on their phones.
What we really needed were policy and regulatory changes aimed at the products themselves… especially the algorithms Meta and Tick Tock use to keep kids scrolling to sell more ads and make more money.
A recent study of more than 32,000 adolescents found that about one-third had encountered self-harm content online during the previous month. Among kids who saw it, the most common way it reached them was through algorithmic recommendations—not because they went looking for it.
Another recent study examining actual Instagram and TikTok feeds found that engagement-driven algorithms can create feedback loops, repeatedly serving young people content based on what captures their attention or reflects their emotional state.
That’s why the new multistate settlement with Meta last week has the potential to be an important public health intervention.
And make no mistake: the changes Meta agreed to weren’t voluntary. They were crammed down Zuckerberg’s by state attorneys general and their staff.
The settlement agreement includes Arizona and revolves around the things Meta has done within Facebook and Instagram (mostly Instagram) by intentionally using addictive features and knowingly exposing kids to serious mental health harm… and lying to the public about the safety of its products.
Of course, Meta admits no wrongdoing in the settlement agreement (eyeroll), but they did sign off on some potentially positive public health interventions.
For users under 18, Meta agreed to a default two-hour combined daily limit on Facebook and Instagram, some interruptions during prolonged use, blocking most use from midnight to 6 a.m., fewer notifications overnight & during school hours, hidden ‘like’ and reaction counts, restrictions on cosmetic-procedure filters (not sure what that is), a way to verify age for users and some basic (undefined) protections from harmful content.
That’s good public health policy.
But Don’t Celebrate too Much
Thankfully the settlement sets up an independent auditor with access to relevant raw data, internal documents, communications, systems and personnel. If the auditor finds material weaknesses, Meta is supposed to develop and implement a corrective-action plan.
Without enforcement from the auditor all these interventions become suggestions to Meta – not actual requirements.
The attorney’s general can also go back to court if Meta violates it (assuming the court approves the agreement). Sadly, but not surprisingly, Meta’s lawyers built in some loopholes.
Parents (or clever kids who know how to simulate their parents) can avoid some of the algorithm changes which won’t be picked up by auditors. Some of the restrictions I have above have exceptions.
The big loophole is that Meta can replace certain safety tools with alternatives it “reasonably” expects to work at least as well. And, if a design replacement later performs worse, that alone doesn’t automatically put Meta in breach of the agreement.
There also isn’t a simple automatic fine every time Meta gets caught out of compliance. Enforcement depends on auditors finding problems, attorneys general pursuing them and courts acting.
So don’t expect Meta to suddenly care about mental health. Expect it to test the boundaries of the agreement and get away with whatever they can that makes them money.
But… the settlement gets a core public health principle right:
Change the environment. Change the defaults. Measure whether it works and most importantly, make sure there is a talented and independent auditor examining Meta’s compliance with the agreement.
