In the space of two weeks, two of the largest interventions yet into how phones treat young people were announced. Both arrive as switches. Neither tells you whether the switch worked.
On 27 August, Meta agreed to a settlement worth roughly 18 billion dollars with a bipartisan coalition of state attorneys general in the United States, ending a long-running case over claims it used deceptive tactics to keep teenagers on its platforms. Meta denied the allegations. The settlement, which still requires court approval, commits the company to a long list of changes on Facebook and Instagram.
The list is more specific than most of what has come before. Access restricted by default between midnight and 6am. Push notifications switched off between 10pm and 7am, and during defined school hours. A two hour daily limit, with messaging and some long-form content excluded from the count. An option for teenagers to use a feed that does not draw on their previous activity. An option to turn off autoplay. Age assurance systems intended to find accounts belonging to children under 13. Independent auditing of whether Meta actually complies.
Ten days later, on the other side of the world, the Australian government said it would introduce draft legislation requiring platforms to let users switch off recommendation algorithms altogether. Communications Minister Anika Wells framed it as choice rather than prohibition, noting that many Australians value the algorithm for the enjoyment or the utility it gives them. The proposal sits inside a broader Digital Duty of Care framework, which would shift the burden from users reporting harm after it appears to companies assessing and reducing foreseeable risks before it does.
The reaction was not uniform. Opposition leader Angus Taylor said he was highly sceptical and warned the move could amount to censorship. The Greens have argued personalised feeds should be off unless a user chooses them, rather than on until a user turns them off. That distinction, opt-in against opt-out, is still unresolved, and it matters more than it sounds.
What the two have in common
Read together, these are not really rules about content. They are rules about controls. Both hand people a set of switches and trust them to use them.
That is a meaningful shift, and in most respects a welcome one. It treats people as capable of deciding rather than as passengers. It also avoids the harder argument about what content should exist at all.
But it inherits the problem that has followed every control ever shipped, which is that a switch nobody touches changes nothing.
Zvika Krieger, formerly a director of Meta's responsible innovation team, made this point about the settlement directly: research shows people rarely change their settings from the defaults companies offer, and parents may not understand what the options mean in the first place. That is not a criticism of the settlement. It is a description of how software has always worked.
Australia has recent evidence of its own. The under-16 social media ban took effect in December 2025, and studies since have reported that most Australian teenagers were still online three months later. Whatever one thinks of the policy, the gap between what was legislated and what actually happened on the devices is the whole story.
It is whether the behaviour moved.
The question underneath
A notification curfew from 10pm is a good idea. Whether it means your child is asleep earlier, or simply awake without notifications, is a different question, and the platform cannot tell you the answer. A two hour daily limit that excludes messaging is a real constraint, but the number that matters to a parent is not the one the platform counts. It is the total, across every app, on the whole phone.
An algorithm off switch might make a feed calmer. It might also make it dull enough to close, which would be a genuine win, or dull enough to open something else instead, which would not be. Both look identical from inside the platform. Neither is visible to the company that shipped the switch.
This is the structural limitation of platform-level controls: each company can only see its own app. Nobody is measuring the whole.
Four things we would actually watch
If these changes take effect, and you want to know whether they worked rather than assuming they did, these are the things to look at. None require special tools. All are visible on the phone itself.
Total time, not per-app time
If one app gets harder to use and another absorbs the difference, the total will not move. That substitution is the most common way a control quietly fails.
The hour before sleep
Notification curfews are designed to protect the end of the day. If the phone is still busy at 11pm with notifications off, the curfew addressed the symptom and not the habit.
The first hour after school
Restrictions during school hours push use somewhere. Usually that is straight after the bell, and usually nobody is looking.
Whether a limit gets kept or crossed
A limit crossed every day is not a limit. It is information, and it usually means the limit is in the wrong place rather than that the intent was wrong.
Where we stand
We build ThirdLife, so we have an obvious interest in people paying attention to their own phone data. Take the argument accordingly.
We also have a limitation worth naming. Everything ThirdLife works out stays on the phone it runs on, which means we cannot tell you what a thousand people did after these changes landed. We do not have that data, and by design we never will. What we can do is help one person, or one family, see whether the things they changed made any difference on their own device.
Our view is that the direction of both announcements is right. Controls are better than no controls, and defaults matter enormously. The Australian opt-in against opt-out question is not a technicality. It will determine whether the switch is used by most people or by almost nobody.
But controls are the beginning of the work rather than the end of it. Someone still has to look at what happened afterwards. At the moment, the only place that can be seen whole is the device in your hand.
Sources: Time, 27 August 2026, on the Meta settlement. ABC News and The Conversation, 6 and 7 September 2026, on Australia's Digital Duty of Care draft legislation. Al Jazeera, September 2026, on research into the under-16 ban.
Related reading: Parental controls statistics · Screen time and mental health · Screen time statistics Australia