Not a wrecking ball — a scalpel

How to fix Bill C-34

This is not a campaign to kill online-safety legislation. It is a campaign to pass the right one. The amendments below are concrete, separable, and consistent with everything the government says it wants to achieve. If the government refuses all of them, MPs should vote no — and say why.

Keep — and strengthen

  • The duty to act responsibly ss. 31–41: risk assessment and mitigation, blocking and flagging tools, synthetic-content labelling, a human resource person.
  • The duty to protect children through design ss. 20–21: age-appropriate defaults, recommendation-system safeguards, limits on contact-from-strangers — regulation of the machinery, which reaches the harms a ban cannot.
  • Takedown of child sexual abuse material and non-consensual intimate images with the victim complaint process ss. 43–47, 68.
  • Transparency and researcher access ss. 42, 59–64.
  • The private-messaging exclusion s. 11 and the no-imprisonment rule s. 111.

Remove

  • The under-16 account ban and its age-assurance machinery ss. 26–29. This is the Australia experiment, and Australia has run it for us: majority circumvention by teens, privacy costs for every adult. Strike the division entirely.
  • The pornography age-verification mandate ss. 22–23 in its current form. Parliament examined near-identical site-level age verification in Bill S-210 last Parliament and heard overwhelming expert evidence about its privacy and expression costs. If any age-assurance concept survives committee, it must be rewritten to apply to dedicated pornography services only, with an explicit standard excluding general-purpose platforms — not to any service an operator "suspects" provides access to pornographic content.
  • The upload-scanning regulation power ss. 12(2), 126(1)(i). Mandatory reporting and takedown of child sexual abuse material already exist in law and in this bill. A regulation-making power to compel scanning infrastructure is a standing invitation to scope creep. If Parliament wants upload-screening of known CSAM hashes by large platforms, it should legislate that — narrowly, in primary legislation, with judicial oversight — not hand cabinet a blank technological mandate.

Amend

  • Turn the expression caveats into prohibitions. Sections 22(3) and 27(3) say the Act doesn't require measures that "unreasonably or disproportionately limit users' expression." Adopt the CCLA's fix: operators must be prohibited from implementing such measures, and the Commission directed to enforce that prohibition.
  • Restore Parliament. Any regulation that designates services, sets age-assurance standards, or invokes s. 12(2)-type powers should require affirmative resolution of both Houses, a published Charter assessment, and a sunset with mandatory review. Fifty unreviewable decisions is not oversight cf. ss. 5–8, 27(5), 126–127.
  • Tighten "harmful content." Definitions like "content that foments hatred" import contested human-rights-code standards into a regime enforced by revenue-percentage penalties. Committee should narrow the definitions and build in an over-removal check: require platforms to report lawful-content removal rates, and empower users to appeal takedowns.
  • Due process at the Commission. A regulator that can compel evidence like a superior court s. 73 while "not bound by any legal or technical rules of evidence" s. 74 and can fine 3% of global revenue s. 88 needs procedural guarantees to match: clear evidentiary standards in penalty proceedings and full appeal rights.

The honest comparison table

Two paths to the same goal
Question Duty of care (keep) Age mandates (remove)
What it regulates Platform design, systems, processes Who may access the platform
Who bears the burden The companies that profit from the risk Every user, starting with the most vulnerable
Does it reach algorithmic harm? Yes — directly (risk-mitigation duties) No — untouched for everyone 16+
Evidence it works Early but positive (UK/EU systemic-duty regimes) Australia: 60%+ of banned teens still have access
Charter exposure Manageable with tightened definitions Severe (s. 2(b): anonymity, access, minors' rights)
ICCPR / CRC compliance Aligned with General Comment No. 25 (safety by design) Collides with General Comment No. 34, para. 43 (generic bans)
New data risks created None for users ID/biometric verification layer across the internet

"But most Canadians support a ban."

Polling does suggest broad public sympathy for an under-16 minimum age, and honesty requires saying so. Three answers. First, rights protections exist precisely for the moments when restrictions are popular; that is the entire point of constitutionalizing them. Second, support measured before the costs are visible is soft — Australians supported their ban too, before adults met the face-scans and parents discovered their teens were on VPNs. Third, the same polls show what people actually want: children safer online. The duty of care delivers that. The ban delivers the appearance of it.

What happens next — and where you fit

  1. Second reading (expected fall 2026): MPs debate and vote on the principle. Constituent mail before this vote shapes caucus positions.
  2. Committee study: clause-by-clause review, expert witnesses, amendments. This is where ss. 22–29 can be removed or rewritten.
  3. Report stage and third reading, then the Senate — which, in the last Parliament, showed willingness to slow flawed internet legislation.

The system responds to volume and to specificity. A letter that says "remove sections 26 to 29, keep the duty of care" is worth a hundred that say "I'm concerned."

Send that letter now