Canada Canada Online Streaming Act CRTC C-11

Canada's C-22 Caps Retention at Six Months but Leaves Encryption and Scope to Regulation, and the Senate Should Close Those Gaps

Bill C-22's safeguards are real, but its 'systemic vulnerability' test and delegated powers leave encryption exposed and put Canada's EU adequacy status at risk.

Bill C-22 by the numbers People of Internet Research · Canada 6 months Retention cap in House text Ceiling on regulations requiring m… 2 years Maximum ministerial order duration Orders need Intelligence Commissio… 16 Civil society letter signatories European groups asking the EU to a… 18 June House passage date Bill passed third reading in the H… peopleofinternet.com
Bill C-22 by the numbers People of Internet Research · Canada 6 months Retention cap in House text 2 years Maximum ministerial orde… 16 Civil society letter signatori… 18 June House passage date peopleofinternet.com

Key Takeaways

Canada's Lawful Access Act, 2026 (Bill C-22) is now in the Senate. According to Parliament's LEGISinfo record, it passed the House of Commons on 18 June 2026, completed Senate second reading on 22 October, and was before the Standing Senate Committee on National Security, Defence and Veterans Affairs on 2 November. The Senate is the last place the bill can be amended before it becomes law.

The strongest case for the bill

Police and intelligence agencies have a real problem. Investigators with a lawful warrant often find that providers kept no usable records, or that the provider is a foreign service with no obligation to cooperate. A state that can obtain a warrant but cannot get the data it covers has a gap in its justice system. C-22 also contains more safeguards than earlier "lawful access" proposals. In the House-passed text, the Governor in Council may require retention of categories of metadata only "for reasonable periods of time not exceeding six months" (s. 5(2)(d) of the new Supporting Authorized Access to Information Act). Section 5(4) bars regulations that would require retention of content, web browsing history or social media activity. Ministerial orders last at most two years and are not valid until the Intelligence Commissioner approves them (s. 7). Section 2(4) states that no obligation under the Act should be construed as compelling a provider to decrypt data it cannot decrypt.

These limits matter, and a fair critique has to start from them.

Where the safeguards thin out

The problem is how much the bill leaves to regulation and ministerial order. The six-month figure is a ceiling on what cabinet may require, not a retention duty written into the statute. Which categories of metadata are retained, and which providers count as "core providers," will be set by regulation after Parliament has finished with the bill. Section 5(4.1) requires cabinet to be satisfied that a category is "essential" to investigations. That is an executive test, not a legislative one. EFF's 22 October analysis describes the retention mandate as running "up to a year," which does not match the House-passed text I reviewed. A mismatch like that between a bill and its critics' descriptions shows how hard it is to judge a regime whose key parameters sit in future regulations.

The encryption protections are narrower than they first appear. The decryption carve-out in s. 2(4) applies only to data the provider cannot decrypt. It excludes cases where "the encryption was provided by the electronic service provider and the provider possesses the information necessary to decrypt." That covers many cloud and storage services.

The "systemic vulnerability" test is also less protective than the name suggests. The statute defines it as a vulnerability that creates "a credible risk" that secure information could be accessed by someone with no authority to do so. It then excludes risk that "relates only to information related to persons with respect to whom a warrant, order or other authority... applies." On that wording, a weakness built to reach warrant targets may not count as "systemic." EFF reads the provision as allowing the Minister to "order companies to create a backdoor to their services for law enforcement, as long as it doesn't lead to a 'systemic vulnerability.'" Access Now's open letter says the bill "does not prohibit orders that defeat end-to-end encryption."

Security engineers will say that a vulnerability limited to chosen targets is not a stable category. An access mechanism has to be built, maintained and protected from misuse, and the people who build it cannot choose who finds it. Ministerial orders can also contain "any provision that may be contained in a regulation," whether or not the provider is a core provider (s. 7(1)). Providers get a chance to make representations (s. 8), but a small developer with a few thousand Canadian users has far fewer resources for that than a large platform.

The innovation and trade cost

Pro-innovation regulation aims to make compliance costs predictable. C-22 reaches an "unidentified class" of electronic service providers, which the Access Now letter lists as including "messaging apps, cloud storage, connected vehicles, camera networks, and health trackers." The same letter says the bill reaches any provider whose services are used by people in Canada. A start-up deciding where to launch a privacy-preserving product has to guess what the regulations will require and whether a ministerial order could force a redesign.

The EU dimension adds a second cost. The Access Now letter, signed by 16 civil society organisations and addressed to Commission leaders, says the bill's metadata retention "is irreconcilable with EU law and will jeopardise Canada's adequacy status." Access Now's accompanying press release asks the Commission to start a review under Article 45(4) GDPR and, if safeguards cannot be secured, to use its Article 45(5) powers to suspend adequacy. These are advocacy positions, not Commission findings, and the Commission has not acted on them. But adequacy underpins data flows between Canadian and European firms. A fight over it would be costly for Canadian companies that did nothing to cause it.

What the Senate should do

Rejecting the bill outright is not the proportionate answer, given the real investigative gap and the safeguards already in the text. A Senate committee that is serious about fixing the bill has several concrete amendments available:

The bill's backers have already accepted the principle that lawful access must have limits, and they wrote some of those limits into the text. The Senate can finish the job by making the limits depend on the statute rather than on the discretion of whoever holds the Minister's office. The point of those amendments is to keep Canadians secure from criminals and from the risks that surveillance mandates create.

Sources & Citations

  1. Parliament of Canada: Bill C-22 (45-1), Third Reading text, as passed by the House of Commons June 18, 2026
  2. Bill C-22 text as passed by the House (third reading, XML)
  3. EFF: Canadians Are Running Out Of Time To Oppose The Dangerous C-22 Surveillance Bill (22 Oct 2026)
  4. Access Now: Open letter from European civil society organisations on Bill C-22 (14 Sep 2026)
  5. Access Now: The EU must act now to protect privacy and encryption from Bill C-22