Singapore's Prime Minister has put social media platforms on notice. In his National Day Rally speech on 23 August 2026, Lawrence Wong said that on age, "self-declaration is just not good enough", and that platforms must have "robust and reliable checks" so the minimum age is actually enforced, as Mothership reported. For teens aged 13 and above, he called making social media "safer by design" the immediate priority. If safeguards stay inadequate, the government will consider "raising the minimum age for accessing the platform beyond 13."
The direction is sensible. The last step, a higher age floor, is the part to watch.
The strongest case for the government
The case for acting is serious. A Ministry of Health expert panel report from July, summarised in The Star's coverage, found that one in three users aged 15 to 17 spends more than three hours a day on social media. It also found that one in six users aged 10 to 24 shows signs of problematic use. The panel flagged infinite scroll, autoplay and visible like counts.
The current system also plainly does not work. A minimum age of 13 that relies on a typed-in birth year is a rule nobody enforces. A government that says so is being honest, not authoritarian. Singapore has also built the pieces for a layered approach, which most democracies lack.
What Singapore has already built
Three pieces matter here.
- App stores. IMDA issued its Code of Practice for Online Safety for App Distribution Services in January 2025. Per Allen & Gledhill's summary, it took effect on 31 March 2025. Designated app stores must have age assurance, covering both estimation and verification, and must file an implementation plan with IMDA. The plan must explain how they will identify children "with reasonable accuracy" while complying with data protection rules. The first target is keeping children away from the highest age-rated apps, such as 18+ titles.
- Harm remedies. The Online Safety Commission began operating on 29 June 2026 under the Online Safety (Relief and Accountability) Act 2025. According to the Ministry of Law, its first phase covers five harms: online harassment, doxxing, online stalking, intimate image abuse and image-based child abuse. For harassment and stalking, victims report to the platform first. They can go to the Commission if the platform has not responded adequately within 24 hours. The Commissioner can direct that content be disabled or an account restricted, and ignoring a direction is a criminal offence. Decisions can be reconsidered within 14 days and then appealed to a panel.
- Social media next. In a 6 May 2026 parliamentary reply, MDDI said it plans to extend age assurance requirements to Designated Social Media Services. It said it is studying features such as direct messaging and autoplay. It also said it is discussing expectations with those services and will consult the public.
This is a reasonable sequence. Harms get a fast, appealable remedy, the app store gets a gatekeeping duty, and the platforms get a conversation before a mandate. The Rally speech adds a deadline to that conversation. Minister Josephine Teo has said legislative changes will be tabled in early 2027.
Where the risks are
Age checks are not free. The Star reports the options under discussion include government ID, credit card data, face scanning or analysis of online usage. Each creates a data-collection problem. A platform that must prove everyone's age ends up holding sensitive data on everyone. That data is exactly what the Personal Data Protection Act is supposed to minimise. Singapore's own app store code recognises this by tying age assurance to data protection compliance. The social media rules should copy that language. They should also require that age signals be checked once and not stored. App store-level assurance can serve as a shared signal, so that a teenager is not asked to scan their face for every app.
Accuracy cuts both ways. Age estimation errs in both directions. Too lax, and children get through. Too strict, and adults are locked out or pushed to less regulated services. "Reasonable accuracy", the standard in the app store code, is the right bar. It should come with published error rates and an appeal route for people wrongly refused.
The age-floor threat needs a limit. Teo's remarks, as The Star reports them, describe a higher minimum age of 18 for any platform that cannot or will not introduce the safeguards. As a deterrent, that is understandable. As a policy it is blunt. It cuts off a 15-year-old from a school group, a hobby community or a family chat, along with the feeds that cause harm. The Rally speech frames it as a consequence for platforms that fall short, and it should be applied that way. The tests should be measured, published and applied to each platform, with a chance to remediate before any age floor is raised.
Design rules beat access bans. Teo has said features that prolong use, such as autoplay and infinite scroll, would be off by default for young users, along with stranger contact and some algorithmic feeds. These target the mechanism behind the harm, not the speech itself. They are easier to defend on free expression grounds than a ban on teenagers' access, because teens keep the ability to read, post and organise.
What a proportionate rulebook looks like
First, publish the consultation paper and let researchers and civil society see the age assurance standards before they harden into law. MDDI has said it will consult, and a consultation held before the early-2027 bill would honour that.
Second, write data minimisation into the statute: no retention of identity documents or face scans after the check, and no reuse for advertising.
Third, require platforms to report on what the safeguards achieved, such as under-18 time spent, stranger-contact rates and appeal outcomes. If the 13-to-17 measures work, the age floor never needs to move.
Fourth, build in review. The Online Safety Commission is a few months old and has covered five harms so far. Singapore will learn from how it performs, and the age-check rules should be written so they can be adjusted in light of that.
Singapore has so far avoided the instinct to ban first and design later. Its layered, regulator-led approach is more likely to survive contact with real users than a flat age ban. The Rally speech was right that self-declaration fails. Whether the response stays proportionate will depend on the 2027 bill, specifically how it handles data minimisation and when it allows the age floor to be raised.