Research Review: What Evidence Supports Australia's Social Media Age of 16?

Australia's social-media minimum-age law is frequently described as evidence-led. That phrase needs unpacking.

There is evidence that some young people experience serious harms through social media. There is evidence that voluntary age limits were inconsistently enforced. There is evidence that particular activities, content and design features can interfere with sleep, intensify appearance pressure, expose children to unwanted contact or make harmful material easier to encounter.

But those findings do not automatically prove that preventing under-16s from holding accounts will improve mental health. That is a separate question—and it is only beginning to be tested.

This review examines the evidence behind Australia's decision, the earliest implementation findings and the important gaps that remain.

What policy is being evaluated?

Australia's Online Safety Amendment (Social Media Minimum Age) Act 2024 (https://www.legislation.gov.au/C2024A00127/latest/text) requires designated platforms to take reasonable steps to prevent Australians under 16 from creating or keeping accounts. The obligation commenced on 10 December 2025.

Children and parents are not penalised for access. The regulatory duty sits with platforms.

That matters when interpreting the policy. This is not simply a recommendation that families delay social media. It is a structural intervention intended to change platform behaviour and make delayed access a shared social norm.

What research question did Australia answer?

The legislation did not answer “At what exact age does social media become safe?” Research cannot provide one universal number because social media is not one exposure and children do not respond uniformly.

A teenager messaging a close friend, watching recommended self-harm videos, posting artwork, being harassed by strangers and scrolling an image-heavy feed are doing different things. Effects depend on the child, the activity, the content, the surrounding relationships and what online activity displaces.

The policy instead answered a normative question: when the evidence is uncertain but possible harm is serious and widespread, who should carry the burden of protection?

Australia's answer was that platforms should carry more of it.

Why age 13 was a weak foundation

The OECD's 2025 analysis of 50 services used by children (https://doi.org/10.1787/a19853ab-en) helps explain why governments became dissatisfied with the existing system.

Thirty-four services stated a firm minimum age. Yet only two systematically assured a user's age when an account was created, regardless of location. Many relied on a date of birth entered by the user or introduced checks only for particular features, jurisdictions or suspicious accounts.

The report also found that minimum-age language could be confusing and that service thresholds were often shaped by privacy or contractual considerations. The familiar age of 13 was not a scientific conclusion that commercial social media becomes developmentally safe at that point.

The policy case for change therefore included an enforcement problem as well as a health concern. A minimum age that users could pass by entering a different birth year provided limited protection.

What does the mental-health research show?

The strongest defensible conclusion is mixed and specific: social media can be beneficial, neutral or harmful, and average population effects can conceal severe experiences for smaller groups.

In a major UK study, Orben and colleagues (https://doi.org/10.1038/s41467-022-29296-3) examined age-varying associations between self-estimated social-media use and life satisfaction. The relationships were small, differed across age and sex, and ran in both directions. Higher use sometimes predicted lower later life satisfaction, while lower life satisfaction also predicted later increases in use.

That study is important because it rejects a one-size-fits-all model. It does not identify a universal danger age or prove that social media caused the observed changes.

A US longitudinal cohort study by Riehm and colleagues (https://doi.org/10.1001/jamapsychiatry.2019.2325) found associations between time spent on social media and later internalising or combined internalising and externalising problems after adjustment. Like much of this field, it relied on reported use and could not fully resolve unmeasured confounding or distinguish different platforms, content and activities.

Other longitudinal work archived for this review similarly suggests that user patterns and context matter. Prospective systematic-review evidence identifies sleep as a plausible pathway, although device use is broader than social media and measurements vary substantially.

Research therefore supports concern about cyberbullying, unwanted contact, harmful recommendations, appearance comparison, compulsive patterns and sleep disruption. It also documents benefits including friendship, creativity, information and support.

What it does not establish is that a uniform account restriction will produce a uniform benefit.

Why choose 16?

Sixteen can be defended as a clear policy boundary that gives younger adolescents more time before entering full commercial social-media environments. Developmental evidence supports the general proposition that sensitivity, self-regulation and social context change through adolescence.

It does not demonstrate that every 15-year-old is unready or every 16-year-old is ready.

The most accurate description is therefore:

Sixteen is a legal threshold informed by development, social convention and precaution—not a biological discovery.

The Australian Senate inquiry (https://www.aph.gov.au/Parliamentary_Business/Committees/Senate/Environment_and_Communications/SocialMediaMinimumAge/Report) records the government's child-safety rationale as well as objections involving implementation, participation and privacy. The Parliamentary Joint Committee on Human Rights (https://www.aph.gov.au/-/media/Committees/Senate/committee/humanrights_ctte/reports/2024/Report_11/Report_11_of_2024.pdf) scrutinised proportionality, privacy and freedom of expression.

Those are not peripheral concerns. They are part of the evidence needed to judge whether the intervention is proportionate.

What does the earliest Australian study show?

A 2026 observational study published in The BMJ (https://www.bmj.com/content/393/bmj-2026-363695) followed 436 Australian adolescents around implementation.

At follow-up, more than 85% of participating under-16s reported using platforms covered by the law. Some retained their own accounts, used false ages or described other forms of circumvention. The analysis found insufficient evidence of an immediate discontinuity in social-media use at age 16.

This is useful early evidence of implementation difficulty. It is not a definitive evaluation of the law.

The study was underpowered, depended on self-report and occurred during incomplete implementation. Its design also faced limitations relevant to a regression-discontinuity analysis. The authors did not establish whether a more mature implementation would change later cohorts or improve health outcomes.

Separately, eSafety's early evaluation (https://www.esafety.gov.au/newsroom/media-releases/early-insights-from-esafetys-comprehensive-evaluation-project) reported that the proportion of under-16s holding social-media accounts fell from 52.4% to 42.1% in its study.

These results should not be forced into a simple contradiction. The studies used different samples, methods and outcomes.

Accounts removed, continued use, harmful exposure and wellbeing are four different outcomes.

What would convincing evidence look like?

A robust evaluation should follow young people over time and examine more than account ownership.

It should ask whether children:

- experience less cyberbullying, coercion and unwanted contact;

- encounter less self-harm, eating-disorder or other harmful material;

- sleep better;

- experience improved distress and wellbeing measures;

- retain healthy connection, support and access to information.

It should also examine displacement to excluded or less regulated services, errors in age assurance, privacy consequences and unequal effects across groups.

The cohort question is especially important. A 15-year-old losing an established account may respond differently from a child who grows up expecting social-media access to begin later. Short-term studies of existing users cannot settle that question.

How does international policy help test the rationale?

Different jurisdictions are pursuing different policy theories.

Australia's model focuses on account access. The United Kingdom combines existing child-safety duties with an announced under-16 service restriction intended for 2027. The European Union's operative Digital Services Act model emphasises risk assessment and safety by design: safer defaults, recommender-system changes, limits on unwanted contact and proportionate age assurance.

The European Parliament supports a digital majority age of 16 with parental consent between 13 and 16, but that is a non-binding position rather than an EU-wide ban. New Zealand's Australian-style proposal remains a member's bill. France illustrates the possibility that courts may reject blanket restrictions when they are insufficiently tailored to privacy, expression and family context.

These differences create a natural policy comparison. If several models are evaluated well, researchers may eventually learn whether account exclusion, safer design or a combination provides the greatest benefit with the fewest unintended harms.

Little Dopes assessment

Australia had a defensible reason to act. Existing age gates were weak, children faced real harms and parents were being asked to manage systems they did not design and could not audit.

But a defensible reason to act is not evidence that the chosen intervention has worked.

The fairest current conclusion is that Australia has begun a large social experiment grounded in precaution and platform accountability. Early evidence shows that implementation is difficult. It does not yet establish the long-term effect on harm, sleep or wellbeing.

The standard for success should not be how many accounts disappear from a dashboard. It should be whether childhood becomes safer without unnecessarily cutting young people off from connection and support.

Sources reviewed

- Online Safety Amendment (Social Media Minimum Age) Act 2024 (Cth).

- Australian Senate Environment and Communications Legislation Committee report on the Bill.

- Parliamentary Joint Committee on Human Rights, Human rights scrutiny report 11 of 2024.

- OECD (2025), Age Assurance Practices of 50 Online Services Used by Children.

- Orben et al. (2022), Windows of developmental sensitivity to social media.

- Riehm et al. (2019), Associations Between Time Spent Using Social Media and Internalizing and Externalizing Problems Among US Youth.

- Full-text UK longitudinal social-media studies and the prospective device-use/sleep systematic review listed in the Little Dopes Month 3 primary-research register.

- BMJ (2026), early observational evaluation of the Australian law.

- Australian longitudinal evaluation protocol.

- eSafety Commissioner (2026), early evaluation insights.

- European Commission DSA guidelines on protection of minors.

- European Parliamentary Research Service (2026), comparative briefing.

- UK Government (2026), Growing Up in the Online World response.

- New Zealand Social Media (Age-Restricted Users) Bill.

Legal and policy statuses were checked on 23 August 2026 and should be refreshed immediately before publication.