What section 2(b) protects
Canadian courts read s. 2(b) broadly. Since Irwin Toy v. Quebec (A.G.), [1989] 1 S.C.R. 927, any activity that attempts to convey meaning is protected expression, and the guarantee covers listeners as well as speakers — the right to receive information is part of the right. Posting to a platform, reading a feed, joining a community: all of it is expressive activity at the core of 2(b).
Two further doctrines matter here:
- Anonymity. In R. v. Spencer, 2014 SCC 43, the Supreme Court recognized that informational privacy includes anonymity — that "anonymity may, depending on the totality of the circumstances, be the foundation of a privacy interest." People speak, search, organize, leave abusive homes, question their faith and explore their identity online because they can do so without first identifying themselves. A regime that conditions access to the main venues of public expression on age-proof — in practice, identity-proof or biometric proof — burdens that interest for every user, not just children.
- Minors hold Charter rights. Children are not rights-free until their 16th birthday. In A.B. v. Bragg Communications, 2012 SCC 46, the Court protected a teenager's ability to pursue her own online-harassment case anonymously, recognizing both the privacy and expression interests of young people. The under-16 ban does not tailor children's online experience; it removes them from the medium entirely — from the place where their generation's civic, creative and political life happens.
How C-34's age provisions burden the right
- For adults: sections 22 and 27 condition account-holding on passing an age check. The state does not directly collect your ID — but state law compels platforms to demand it, and a burden imposed through conscripted intermediaries is still a state-imposed burden. The practical effect: the end of anonymous participation in designated public forums.
- For under-16s: section 27 is a complete prohibition on holding an account on designated services — a ban on access to entire media of communication for a class of citizens, enforced not against the children but through the platforms.
- For everyone: penalty-backed duties to mitigate vaguely-bounded "harmful content" invite over-removal of lawful speech (collateral censorship). The risk is structural: platforms answer to the regulator, not to the users whose posts disappear.
The section 1 test, honestly applied
Charter rights are not absolute. Section 1 permits "such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society," assessed under the Oakes framework. Here is how that analysis would actually run — including where the government wins.
Where the government wins
Pressing and substantial objective? Yes — easily. Protecting children from online harms is precisely the kind of objective courts accept, and the harms are not invented: exploitation, sextortion, algorithmic amplification of self-harm content. The Supreme Court has upheld child-protection limits on expression before — Irwin Toy (advertising directed at children), R. v. Sharpe, 2001 SCC 2 (child pornography offences). Anyone who tells you this bill is obviously, entirely unconstitutional is overstating the case. The fight is not about the objective. It is about the means.
Where the age mandates fail
- Rational connection — strained. A measure must actually further its objective. Australia's identical ban has been in force since December 2025; within months, a Molly Rose Foundation survey found over 60% of teens who had accounts still had access, while researchers documented displacement to VPNs and less-moderated spaces. A ban that determined teenagers route around, but every adult pays for, has a connection problem on the evidence.
- Minimal impairment — the fatal step. The government must show it chose a means that impairs the right as little as reasonably possible. C-34 cannot, because the less-impairing alternative is printed in the same bill: the duty of care. Safety-by-design duties, blocking and flagging tools, takedown of the worst content, transparency and audits — sections 20–21 and 31–47 — pursue the same objective while restricting no one's access and identifying no one. A statute that contains its own less-restrictive alternative makes the minimal-impairment argument for the challenger.
- Proportionality of effects — the toll is universal. The deleterious effects fall on every Canadian (identification as a condition of expression; a new class of breach-prone databases or biometric checks), on the most vulnerable first — abuse survivors, whistleblowers, closeted teens in hostile homes, undocumented people — and on under-16s completely. Weigh that against a benefit Australia's experience shows to be substantially illusory, and the balance tips.
The vagueness overlay
"Prescribed by law" demands intelligible standards. C-34 delegates the operative content of the limits — which platforms, which measures, what counts as "adequate" — to roughly fifty future cabinet and Commission decisions. Courts have tolerated delegation, but a limit on expression whose scope Parliament itself cannot describe at the moment of voting sits uneasily with the rule-of-law foundation of s. 1, and with basic democratic accountability for speech regulation.
The strongest counter-arguments, stated fairly
A site about rigour should model it. Here is the other side's best case, and our answers.
"The Charter permits reasonable limits, and protecting kids is reasonable."
Agreed on both halves — which is why our demand is amendment, not abandonment. The duty-of-care core is the reasonable limit. The question Oakes asks is not "is the goal good?" but "did you choose means that impair rights as little as reasonably possible?" An ID-checked internet and a categorical ban are not those means while safety-by-design sits unexploited in the same statute.
"The bill has privacy safeguards — data must be destroyed after verification."
It does, and they are better than Australia's first draft. But destruction-after-use governs the verification record, not the fact of verification: the anonymity of access is gone either way. Safeguards mitigate the breach risk; they do not restore the right to read and speak without first proving who you are. And a safeguard that lives alongside s. 27(2)(e) — "any other requirement specified in the regulations" — is a safeguard whose real content is, again, cabinet's to write.
"Age checks are normal — we card teenagers for alcohol."
Alcohol is a commodity; expression is a constitutional right. We card at the liquor store because there is no Charter right to buy beer. The closer analogy: requiring ID to enter a library, join a protest, or buy a newspaper. We have never accepted that, and the Supreme Court's recognition in Spencer that anonymity can ground a privacy interest explains why. There is also a scale difference: carding happens at one till, once; age assurance happens across every designated platform, continuously, generating records as it goes.
"Social media demonstrably harms teens. Doing nothing is not neutral."
The harms are real and we do not minimize them — though the research community remains genuinely divided on causation versus correlation. But "do nothing" is not on the table. The choice before Parliament is between two somethings: a duty-of-care regime that forces platforms to fix the machinery (recommendation systems, dark patterns, contact-from-strangers defaults), or a ban that Australia's data shows teenagers evade — pushing the most at-risk kids into spaces with no safety tooling at all, while their better-supervised peers lose the supports, communities and information the UN's children's-rights framework says they are entitled to. If you believe the harms, you should want the remedy that actually reaches the machinery.
"Parliament needs regulatory flexibility; technology moves fast."
Flexibility about technical standards is normal. Flexibility about the scope of a speech restriction — which platforms, which Canadians, which measures — is different in kind. Parliament can keep pace through sunset clauses, mandatory parliamentary review of designations, and affirmative-resolution procedures for speech-affecting regulations. "Trust cabinet" is not a constitutional standard.
The bottom line
C-34's objective would survive a Charter challenge. Its age mandates, on the Australian evidence and against the less-restrictive alternative printed in its own pages, should not. But Charter litigation takes years, costs millions, and happens after the infrastructure is built. The cheaper, faster, more democratic fix is for Parliament to amend the bill now — which is where you come in.