Meta has agreed to pay up to $18bn to settle claims that Facebook and Instagram failed to protect children and were designed in ways that encouraged addictive use.
Meta has not admitted wrongdoing. As part of the settlement, however, the company has agreed to introduce default daily limits, nighttime restrictions, stronger age checks, and additional controls for parents.
Now, on the surface, I think most of that makes sense.
Parents are responsible for supervising their children. They should be able to determine how much time their children spend on social media, what they are exposed to, and when they should put the phone down.
But there is another part of it that I think we have to look at.
A parent might be responsible for the child, but the parent does not control the platform. They do not control the algorithm, the notifications, the recommended content, or any of the other features that are designed to keep the child engaged.
So, for example, a parent can tell a child that they have had enough screen time and need to put the phone away. At the same time, the child is receiving notifications, seeing that their friends are online, and being shown more of whatever content the platform has already learned will hold their attention.
The parent is trying to manage the behaviour from the outside. Meta controls the environment producing the behaviour from the inside.
And I think that is where the question of responsibility becomes more nuanced.
I do not think it is reasonable to say that parents have no responsibility. They obviously do. But I also do not think companies can design these systems, benefit financially from keeping children engaged, and then frame child safety mainly as something parents have to manage through settings.
If the risk is partly being created by the design of the platform, then some of the protection also has to be built into the design of the platform.
Children are not simply smaller adults
Another part of the discussion is the idea that children need to learn to use social media responsibly.
I agree with that. Children have to develop self-control. They have to learn that they cannot respond to every notification, follow every trend, or stay online for as long as they want.
Now, that being said, they are still children.
Their ability to control impulses, think about long-term consequences, and ignore immediate social rewards is still developing. Teenagers are also particularly sensitive to approval, rejection, and what their friends are doing.
Social media did not create those developmental characteristics. But platforms can be designed in ways that take advantage of them.
Infinite scrolling removes a natural place to stop. Notifications create a reason to return. Likes make approval visible. Recommendation systems learn what keeps a particular person watching and then provide more of it.
So we place a child inside that environment and then tell the child to exercise better judgment.
That is part of what they have to learn, yes. But in child safeguarding, we do not normally manage a foreseeable risk only by telling a child to become less vulnerable. We also expect the adults and institutions controlling the environment to make it safer.
I do not see why a digital environment should be treated differently.
The importance of default settings
One part of the settlement that I think is genuinely important is the use of default limits.
An optional safety control helps the parent who knows that it exists, understands how it works, and remembers to activate it. A default changes what happens before the parent does anything.
That is a meaningful difference.
The platform should not start with unrestricted access and then place the safety controls somewhere in a settings menu for parents to find.
But the settlement also has limitations. The restrictions do not apply to WhatsApp or Meta's other direct-messaging services. Florida and New Mexico did not join the agreement. Meta has also connected some tighter protections to YouTube and TikTok accepting similar measures.
Meta's position is that teenagers move between several different apps and, therefore, one company cannot solve an industry-wide problem on its own.
There is some truth in that. If Instagram becomes harder to use at night, a child may simply move to TikTok, YouTube, or another platform.
But I think that is an argument for consistent standards across the industry. It should not become a reason for each company to wait on the others.
The business model is the more difficult issue
The strongest criticism in the Financial Times article came from Katharina Kopp of the Center for Digital Democracy.
Her point, essentially, was that the settlement changes particular features without changing the business engine behind them.
That is the part I keep coming back to.
Meta earns most of its money from advertising. The longer people remain engaged, the more opportunities the company has to show advertisements, gather information, and improve its recommendations.
So even with a daily limit or a nighttime block, the wider system is still rewarded for holding attention.
This does not mean that every person working at Meta is consciously trying to harm children. I do not think the argument needs to go that far.
The problem is the incentive. If a system makes money from attention, it will continue finding better ways to capture attention unless something outside the system places limits on it.
And that is why changing one feature at a time may not be sufficient. By the time courts or policymakers fully understand one method of encouraging engagement, the technology may have already moved on to another.
Will $18bn actually change anything?
Eighteen billion dollars is an enormous amount of money. It is one of the largest corporate settlements in US history.
At the same time, Meta earned approximately $60bn in net income last year. The settlement payments will be distributed over ten years, and the company has said that it does not expect them to have a significant effect on its quarterly results.
So I think there is a reasonable question about whether the payment is a genuine deterrent or simply a very large cost of doing business.
That does not make the settlement meaningless. The changes to the platforms may ultimately matter more than the money.
But we will have to look at what happens after the announcement. Do the controls actually reduce harmful use? Can children easily bypass them? Will Meta provide evidence that the measures work? And what happens if those protections are weakened later?
Those questions are not as dramatic as an $18bn headline, but they will tell us whether children are actually safer.
A systems problem cannot be managed only by individual families
Parents will always have responsibilities. We cannot outsource parenting to Meta, the government, or anyone else.
But I also do not think platforms can outsource child safety to parents.
Parents are being asked to manage individually something that the platforms are creating at a systems level.
If a company designs the environment, controls how it operates, and benefits financially from the way children behave inside it, then safety cannot be only an optional tool that parents have to locate and activate.
It has to be part of the design from the beginning.
Perhaps that is the more important shift represented by this settlement. Instead of asking only why a parent did not stop a child from using the platform, we are beginning to ask why the platform was allowed to operate that way in the first place.