Australia has already taken one of the strongest positions in the world on children's access to social media.
Since December 2025, age-restricted platforms have been required to take reasonable steps to prevent Australians under 16 from creating or keeping accounts.
Now the country is considering another step.
The Financial Times reports that Australia's proposed “my feed, my way” law would require social-media platforms to allow users to opt out of algorithmically recommended content. Instead, people could choose to see content only from accounts they deliberately follow.
The proposal appears simple. Give people a choice.
But it raises a more uncomfortable question:
Why were families required to opt out of manipulation in the first place?
Parents have responsibilities—but they did not design this environment
Whenever concerns are raised about children and social media, one response usually follows:
Parents should simply stop giving their children phones.
There is some truth in that. Parents must make decisions about when their children receive devices, which platforms they may use and how much time they spend online. Giving a young child unrestricted access to social media and hoping for the best is not a sufficient parenting strategy.
But that cannot be the end of the discussion.
Parents did not create infinite scroll.
Parents did not design recommendation systems that learn which images, emotions and insecurities will keep a child watching.
Parents did not introduce streaks, autoplay, intermittent rewards, engagement notifications or carefully timed prompts encouraging users to return.
Parents did not conduct continuous experiments on millions of users to discover which design changes would hold their attention for a few minutes longer.
Technology companies did.
We should therefore be careful about treating a family's decision to provide a phone as though it gave a company unrestricted permission to compete for a child's attention using every behavioural tool available.
A parent can make a mistake. That does not absolve a corporation of responsibility for deliberately designing an environment that makes the consequences of that mistake more severe.
This is not a fair contest
We often describe social-media use as though it were a contest between a child's self-control and a parent's discipline.
It is not.
On one side is a family managing work, school, meals, transport, homework, relationships and the ordinary exhaustion of daily life.
On the other side are some of the wealthiest companies in history, supported by enormous datasets, behavioural research, sophisticated algorithms and teams whose commercial success depends on holding attention.
Then we ask why the parent did not monitor every notification, every video, every account and every recommendation entering the child's feed.
That is not a realistic distribution of responsibility.
Parents should establish boundaries. Children should gradually learn self-regulation. But good public policy recognises when individual families are being asked to solve a structural problem alone.
We do not usually respond to unsafe products by telling every household to become an expert in product engineering. We create reasonable standards for the companies designing and selling them.
Digital products should not be exempt from that principle merely because their risks arrive through a screen.
The algorithm changes the nature of the product
There is an important distinction between social media and algorithmically driven social media.
A child choosing to follow friends, relatives, a football club or a musician is making a decision about what they want to see.
An algorithmic feed makes another decision on the child's behalf: what should come next to keep this particular user engaged?
That difference matters.
The platform is no longer functioning only as a place where people share content. It is actively arranging the user's environment, predicting what will capture attention and learning from every pause, click, replay and reaction.
For a child, that may mean being repeatedly directed towards content involving appearance, dieting, misogyny, self-harm, aggression or other harmful behaviour—not necessarily because the child searched for it, but because the system detected engagement.
The question is therefore not simply whether objectionable content exists online.
The question is whether a commercial system should be permitted to discover a child's vulnerabilities and repeatedly organise content around them.
Australia's under-16 restriction was only one layer
Australia's existing law is frequently described as a social-media ban for children. More precisely, it requires covered platforms to prevent under-16s from holding accounts. Children and parents are not penalised, and publicly accessible content may still be viewed without logging in.
That restriction may give younger children and their families valuable breathing room.
But it does not resolve the underlying design problem.
A child eventually turns 16. The algorithm remains.
Adults are also vulnerable to systems designed to intensify anger, anxiety, comparison and compulsive use. Parents themselves may find their attention fragmented while simultaneously being expected to model healthy technology habits for their children.
This is why the new proposal matters.
Australia is moving beyond the question of who should be allowed onto social media and towards the more fundamental question of how social media should be permitted to operate.
Its proposed legislation would also require companies to protect under-18s from harmful content and addictive features across social media, games, apps and AI chatbots. Regulators could impose substantial penalties where companies breach the rules.
That shifts part of the burden back towards the organisations with the greatest capacity to change the environment.
Opting out is helpful—but the default still matters
Giving users the ability to reject algorithmic recommendations is a worthwhile step.
A chronological or following-only feed could return some intentionality to social-media use. People could decide whom they wish to hear from without allowing the platform to fill every gap with whatever content is most likely to provoke another reaction.
But an opt-out model still deserves scrutiny.
Defaults are powerful. Many users do not change settings they do not understand, cannot easily find or did not know existed. Children are particularly unlikely to study a platform's configuration before using it.
For younger users, the safer option should arguably be the default.
A company should have to justify activating a highly personalised recommendation system for a child—not require the child or parent to locate a switch after exposure has already begun.
If a feature carries foreseeable risks and primarily serves the company's need for engagement, the burden should not rest entirely on families to disable it.
Pro-family policy is not anti-technology
None of this requires us to pretend that technology has no value.
Young people use digital platforms to communicate, learn, create, discover communities and access support. Social media can connect families across countries and give isolated young people access to information they may not find locally.
The objective should not be to frighten families or shame parents.
It should be to make the digital environment more consistent with the interests of the people using it.
Parents still need to delay access where appropriate, establish boundaries, discuss online experiences and model healthier habits. Schools still need to teach digital judgment rather than rely exclusively on blocking websites. Young people still need opportunities to develop self-regulation.
But those responsibilities must exist alongside corporate responsibility.
We cannot continue allowing companies to design increasingly persuasive systems and then explain every harmful outcome as a failure of parenting, discipline or individual choice.
Families did not create this problem by themselves, and they cannot solve it by themselves.
Australia's proposal recognises something that digital policy has avoided for too long: the design of the platform is not neutral.
Parents and children should be able to choose what they want to see. They should not have to continuously defend themselves against systems designed to override that choice.