Canada's Social Media Bill Is Better Than a Ban, but Ottawa’s Rollout Must be Right

Canada is on the verge of barring children under 16 from holding social media accounts, with one significant exception: A platform that can prove it has made itself safe for them can keep its young users. That exception, buried in the middle of Bill C-34, the Safe Social Media Act, is the most important part of the bill, and the one most likely to be wasted. If Canada gets the rollout wrong, it will repeat Australia's sequencing error—a restriction that ran ahead of any mechanism to reward safer design. Whether Bill C-34 rewards safer platforms or simply clears kids off them depends on when the restriction comes into force and how the regulator writes the safeguard standards.

A ban forces platforms to get users under 16 off the service. An exemption tells them to build a service safe enough for those users so they can keep them. That is meant to push effort toward design, defaults, and the conditions children encounter once they are logged in. It’s a carrot-and-stick approach. Culture Minister Marc Miller said as much when tabling the bill, arguing that social media can be made safe by design rather than simply emptied of its youngest users. If the government’s intent is to let design make the service safe, design should be given a real chance to do so before the account restriction forecloses the question.

That chance is structurally unlikely to exist on time, because the two halves of the bill are not built the same way. The Digital Safety Commission will still need commissioners, staff, rules, consultation, final criteria, applications, evidence reviews, and enough monitoring to know whether approved platforms are actually safer. Officials estimate the bill itself could take about a year to pass and the regulator another 18 months to stand up, before any of the drafting, consulting, or monitoring work has even begun. Meanwhile, the account restriction takes effect the day the government says it will, amid political pressure to act sooner rather than later. Officials have said the restriction could be brought into force before the Commission is fully operational.

In that gap, the only compliant move a platform has is to remove under-16 accounts, or else face penalties that can reach 3 percent of global revenue. Safe by design, under the government's own admission, could arrive years after the ban does.

Platforms already face pressure from parents, advertisers, politicians, and litigation, albeit unevenly and incompletely. Bill C-34 will either reinforce that pressure or distort it, and sequencing decides which. Get the order right, and firms at least have a reason to keep working on the problem. Get it wrong, and the fastest path to compliance becomes removing the very users the bill means to protect.

Another danger is that adequate safeguards become a government-mandated blueprint for the product, determining which feeds, recommenders, and product features count as safe for children. Regulators cannot set that standard well because they do not know the safest configuration for every social media service; that configuration differs across products and age groups, and whatever works this year will not necessarily be the best answer three years from now. A Commission that tries to design the service will freeze today’s guesses and spend its limited capacity adjudicating a list of features that are either banned or allowed.

And telling firms which features count as safe will lead them to build to the list rather than assess the risks their younger users face. The better model is the one the bill already reaches for in its duty-to-act-responsibly provisions: holding platforms to risk mitigation and transparency, not to a prescribed set of product features. The exemption criteria should follow that logic, requiring risk assessments, independent evaluation, transparency about what under-16s encounter, age-appropriate defaults, parental tools, and evidence that harm is falling.

Australia is the cautionary case, though not only for the reason usually discussed. The familiar lesson is that many are circumventing the ban. The Molly Rose Foundation's polling found that 61 percent of Australians ages 12 to 15 who held accounts on restricted platforms before the rule still have access to at least one afterward, most of them easily.

The more useful lesson is about where the effort went. Restrictions took effect last December, and within a month platforms had removed access to roughly 4.7 million under-16 accounts, absorbing compliance teams, government oversight capacity, and public attention. Yet half the affected children reported no change in safety. That is the trade a blunt ban makes: Finite capacity is spent proving accounts are gone, with little of it left for the underlying harms, such as bullying, exploitation, and self-harm content, that drove the policy in the first place. A government has only so much capacity to spend on child safety in a year.

On sequencing, Ottawa should not bring the under-16 restriction into force until the Commission exists, has published the criteria, and can assess platform compliance. On substance, the exemption should be flexible, evidence-based, and non-prescriptive. It should set outcomes-based rules for reducing harm, not impose a state-mandated design brief frozen in 2027.

What happens next will depend entirely on timing. Switch on the restriction before the exemption works, and Canada gets a ban with a safety valve that nobody can use yet. Get the sequence right, and platforms get to design for a real standard that might actually improve outcomes for children.