Canada Joins Global Fight to Ban Teens from Social Media

Last Wednesday, the Mark Carney government tabled the Safe Social Media Act, Bill C-34, legislation that would make online services responsible for addressing harmful content on their platforms. The headline is a restriction on social media accounts for children under sixteen. Canada did not invent that idea. Australia did, and Australia has now been running the experiment for six months. The early results are worth examining before anyone declares this bill a solution or a failure.

First, what the bill does. It applies to social media services, including livestreaming and user-uploaded adult content services, and to artificial intelligence chatbot services. It organizes obligations under three duties: a Duty to Act Responsibly, a Duty to Protect Children, and a Duty to Make Certain Content Inaccessible. It targets seven categories of harmful content, including content that sexually victimizes a child, content that induces a child to harm themselves, bullying content, incitement to violence, content that foments hatred, terrorist content, and intimate content shared without consent.

The bill also creates a Digital Safety Commission of Canada to enforce the legislation through audits, compliance orders, and administrative monetary penalties. Platforms will have to submit publicly disclosed Digital Safety Plans, label synthetically generated content, and provide accredited researchers access to certain data.

The under-sixteen restriction sits inside the Duty to Protect Children, and it comes with a release valve: the age restriction would be subject to an exemption process should the regulator determine that a service has implemented sufficient safeguards for children. That single clause may end up being the most consequential design choice in the bill.

But before debating whether Ottawa’s solution will work, it is worth remembering why governments across the democratic world converged on roughly the same approach at roughly the same time.

The legislation arriving in Ottawa, Canberra, London, and Brussels follows years of accumulating evidence about what the platforms knew and when they knew it. In March, juries in two American courtrooms returned verdicts against the companies this bill would regulate. In California, Meta and Google were found negligent for failing to warn users about the addictive design of their products.

In New Mexico, Meta was found liable for failing to protect children from exploitation and abuse on its apps. The New Mexico jury found the company engaged in unconscionable trade practices that took advantage of the vulnerabilities and inexperience of children. It deliberated for two days and awarded the statutory maximum, $375 million (US). The case was built in part on an undercover operation: state investigators created accounts posing as underage users and documented those accounts being flooded with sexual grooming and harmful content almost immediately.

What the discovery process surfaced is the part Canadian parliamentarians should read before the C-34 hearings begin. According to internal communications described at trial and reviewed on the Center for Humane Technology’s podcast Your Undivided Attention, a Meta vice president acknowledged internally in 2018 that the company knew the scale of grooming and solicitation messages reaching underage users.

In 2019, another executive asked Mark Zuckerberg for twenty-four additional staff to study problematic use and build safety tools; the request was declined. A 2020 internal chat described child safety as “an explicit non-goal” for that half of the year. In 2021, safety engineer Arturo Béjar emailed Zuckerberg directly to report that the company was deeply undercounting unwanted sexual advances toward minors and received no response. In 2022, Meta cut roughly 100 positions from its integrity and responsibility teams.

Separate court filings unsealed last November, from a class action brought by American school districts, fill in the research side. In 2020, Meta commissioned Nielsen to study what happened when people deactivated Facebook and Instagram for a week. The results showed users felt less depressed, less anxious, and less lonely and compared themselves to others less. Meta cancelled the research program rather than expand it. The same filings include testimony from Instagram’s former head of safety and well-being that the company maintained a seventeen-strike threshold for accounts reported for sex trafficking: sixteen violations before suspension on the seventeenth. Meta disputes that characterization, and the company is appealing both verdicts.

The individual allegations deserve the caution owed to claims still being contested. The pattern, established by two juries and years of internal documents, requires less hedging. The companies measured the harm to children, declined to act on their own findings, and disclosed none of it.

This record explains C-34’s transparency architecture. The Digital Safety Plans, the compliance audits, and the researcher access provisions all answer the same problem: for fifteen years, the only people who could see inside these systems were the people with a commercial interest in what the data showed. The United States, which never passed a federal law, is extracting accountability one jury at a time. Canada is proposing to build it into statute.

Australia passed the Online Safety Amendment (Social Media Minimum Age) Act in late 2024, and the ban took effect on December 10, 2025. Ten platforms fell under it—Facebook, Instagram, TikTok, YouTube, Reddit, Snapchat, Threads, Twitch, X, and Kick—with fines of up to $49.5 million (AUD) for serious or repeated breaches. Children and parents face no penalty; the obligation sits entirely on the companies. There are no exemptions for parental consent, and platforms whose main purpose is messaging, gaming, education, or health communication fall outside the regime.

Six months in, the record shows two things at once, and both are true.

The platforms complied. By mid-December 2025, age-restricted platforms had removed access to 4.7 million under-sixteen accounts across Australia. The companies that spent a decade insisting age verification was technically impossible managed it within days of a legal deadline backed by a $49.5 million penalty.

The children adapted. A Pureprofile study of 1,025 parents, teachers, and young Australians, released at the six-month mark, found that 78 percent of under-sixteens continue to access prohibited platforms. One survey is one survey, and self-reported access is a loose measure. But the direction is consistent with what Australian researchers have observed since launch: children migrating to platforms not covered by the ban and trying to fool age-verification technology, including by posing for cameras with double chins to look older. The same study found 76 percent of respondents still support the ban.

Australians appear to have concluded that resetting the norm has value even when enforcement leaks. That is the honest reading of the Australian evidence. The ban changed what platforms must do. It did not seal the perimeter, because no age-assurance technology can. Australia’s eSafety Commissioner said as much at launch: success is measured by reduction in harm and by resetting cultural norms, the same standard we apply to every other age-based safety law.

• A PR Hoax Created the Year’s Hottest Rock Band. Imagine What It Can Do in Politics
• I Was Lonely and Let an App Pick My New Friends. Here’s How It Went
• The War Against Misinformation Is Over. The Lies Won

Canada is a late arrival to a broad movement. The United Kingdom took a different route through its Online Safety Act 2023, imposing systemic duties enforced by Ofcom, with mandatory age assurance for adult content in force since July 2025. The British experience carries its own warning: the age verification rollout triggered a 1,400 percent surge in VPN signups, according to Proton VPN data. The UK has since launched a consultation on whether to ban social media for under-sixteens outright.

The European Union is converging on the same conclusions. The European Parliament’s internal market committee has urged a European Union–wide digital minimum age of sixteen for social media, France aims to restrict social media access for users under fifteen from September 2026, Denmark reached agreement in November to block access for under-fifteens, and as of May 2026, twenty-three of twenty-seven EU member states were at least considering national legislation. The under-sixteen threshold is becoming the democratic world’s default setting. Canada arriving in June 2026 is late enough to learn and early enough to matter.

Two features distinguish the Canadian bill from everything above. The first is the exemption pathway. Canada is proposing a conditional line: comply with safeguards the regulator deems sufficient, and the age restriction can lift for your service. Done well, this turns the ban into a design incentive, which is where the real safety gains live. Done poorly, it becomes a lobbying channel through which the largest platforms negotiate their way back to the status quo.

The criteria will be set by regulation after passage, which means the most important decisions in this bill have not been written yet. The corporate record above is the strongest argument for writing those exemption criteria tightly: the platforms seeking exemptions are the same companies that juries found had concealed what they knew.

The second is the AI chatbot framework, and no comparable national legislation contains one. The bill sets out obligations for AI chatbots, public-facing services that can mimic human relationships, including the duty to reduce the risk of harmful content being generated, and requires measures to respond when a user expresses ideas of suicide or self-harm. The context is grim and recent. The February mass school shooting in Tumbler Ridge, British Columbia, led the AI minister to summon OpenAI executives to Ottawa. Australia’s law predates the mainstream arrival of companion chatbots. Canada’s does not, and the drafters noticed. Whatever else happens to this bill, the chatbot provisions will be studied in other capitals.

There is a principle underneath all of this that matters more than any single clause. For close to two decades, the rules governing what Canadian children encounter online have been written in California, in terms-of-service documents no legislature ever voted on, by companies whose internal records show they understood the harm.

Australia decided its Parliament outranks a platform’s user agreement. The UK and the EU reached the same conclusion. C-34 is Canada making that decision for itself: the conditions under which foreign platforms operate in this country are a matter of Canadian law, set by people Canadians can vote out. The bill will be amended, the regulations will be fought over, and the age-assurance technology will be imperfect, because Australia has already shown us it will be.

The important thing to remember is we, the people of Canada, have the power to protect our society from technology that is harmful. Big Tech, take notice.

Adapted from “Canada’s Under-16 Ban Arrives With Six Months of Australian Evidence” by B. E. Rybak (Substack). Reprinted with permission of the author.