The case for a ban, stated fairly
The strongest argument for an Australian-style ban is that voluntary platform safety has failed children for a decade. Parents face products engineered for engagement, and individual families cannot out-negotiate them. A hard legal age line gives parents a social norm to point to. It also shifts the burden from the child to the company.
Australia's law, which took effect on 10 December 2025, shows a ban can move platform behaviour at scale. A joint release from Prime Minister Anthony Albanese and Communications Minister Anika Wells on 16 January 2026 said more than 4.7 million under-16 accounts had been deactivated, removed or restricted within days of the law taking effect. The release described this as preliminary analysis from a first tranche of information provided to the eSafety Commissioner.
What Malatsi actually said
According to Biometric Update's 10 September 2026 report, Communications and Digital Technologies Minister Solly Malatsi told Parliament in a Q&A response that the government is not pursuing a blanket policy restricting under-16s from social media. The report quotes BusinessTech on his reasoning: "There is currently no irrefutable indication that a general social media access ban for persons under 16 years old actually prevents them from being digitally active on such platforms."
The same report says Malatsi had earlier, at South Africa's Social Media Summit in July, urged platforms to build stronger internal mechanisms for managing harmful content. I could not retrieve the parliamentary reply itself. The quotations here come from that secondary report.
That is a defensible position, and it is narrower than some coverage suggests. The minister is not saying children are safe online. He is saying the evidence does not yet justify a blunt instrument.
Why the evidence bar matters
The Australian number cited above counts accounts closed. It does not count children protected. A deactivated account is not a child who has stopped using the platform. It tells us what platforms did, not what children experienced afterwards. The government's own release called the figures preliminary and promised continued monitoring of compliance. A minister who wants evidence of effect, rather than evidence of activity, is asking the right question.
The costs of a mandate also fall on everyone, not just children. A legal age gate has to be enforced by verifying age, and verifying age means collecting identity or biometric data from adults and children alike. South Africa already has a statute pulling in the opposite direction. Section 34 of the Protection of Personal Information Act (POPIA) provides that "a responsible party may, subject to section 35, not process personal information concerning a child". Section 35 lifts that prohibition where processing is carried out with the prior consent of a competent person, among other grounds.
A verification mandate would sit awkwardly beside that rule. It would require platforms to build large stores of sensitive identity data in order to keep children out, while the privacy law treats children's data as especially restricted. Good child-safety policy should not create a new honeypot.
Where the minister's position is weak
Rejecting a ban is the easy half. "Platforms should strengthen their own protections" is not a policy unless someone can check whether they did. Voluntary commitments are precisely what the pro-ban camp says has already failed, and they are not wrong to say so.
The credible alternative has to be more demanding than a slogan. It would need at least four things:
- Specific duties. Require risk assessments for features that affect minors, such as recommender systems, direct messaging from strangers and default privacy settings.
- Published evidence. Platforms should report on the effect of their safeguards on children's exposure to harm, not only on the number of accounts removed.
- A named enforcer. South Africa should decide which regulator can demand that information and impose proportionate penalties, rather than leaving responsibility spread across several bodies.
- Use of existing law. POPIA's rules on children's data already give regulators a hook over profiling and data collection aimed at minors.
This is the proportionate course. It targets the design choices that cause harm without requiring identity checks on every South African adult who opens an app.
The speech question
A general ban also restricts something the Constitution protects. Young people use social platforms to find community, information and support, sometimes on topics they cannot raise at home. A blunt ban removes those benefits along with the harms. It is also a poor fit for a country where connectivity is expensive and mobile-first, and where many teenagers may rely on a shared family device. Age-verification rules designed around the assumption that every child has a personal smartphone and a parent's ID are likely to fail the least-connected households first.
None of this means doing nothing. It means choosing interventions whose costs are proportionate to their demonstrated benefits, and being honest that the benefits of a ban are not yet demonstrated.
What to watch
Two tests will show whether this position is serious. The first is whether the department publishes concrete platform obligations, with dates and a named regulator, rather than a further round of summits. The second is whether Australia's enforcement data, as it matures, shows reduced use and reduced harm rather than only closed accounts. If it does, South Africa should revisit the question, and a government that has said it wants evidence will have no principled reason to refuse.
Until then, Malatsi's refusal to copy a headline policy is sound. The measure of the decision is whether the alternative he has chosen is made binding.