A Fourth Force Joins the Network
On June 22, 2026, the Ottawa Police Service Board approved Phase 1 of a facial recognition project that plugs the force into a shared biometric network already running in York, Peel, and Halton regions. The system uses matching algorithms from Idemia Identity and Security Canada, and initial access will be limited to five trained OPS analysts working investigative leads, not live surveillance (Biometric Update). The financial commitment is real but modest by municipal-IT standards: a CA$150,000 deposit against a CA$472,631 one-time implementation cost, rising through annually-increasing maintenance to roughly CA$1.26 million over five years (ID Tech Wire).
What makes this rollout notable isn't the technology — facial matching against mugshot databases is now a fairly standard investigative tool across North American policing — it's that Ottawa becomes the fourth Ontario police service to join this specific shared network absent any dedicated federal statute governing how Canadian police may use facial recognition at all.
The Case for Phase 1
Before dismissing this as reckless expansion, the investigative rationale deserves a fair hearing. Facial recognition can generate leads in cases — child exploitation, violent crime, missing persons — where traditional canvassing fails and time matters. OPS built in real safeguards for a first phase: training includes bias awareness, every match requires independent human review, and the board's own materials state plainly that "a facial recognition result does not confirm identity on its own and is not evidence by itself" (Biometric Update). Limiting access to five analysts during Phase 1, rather than deploying it force-wide, is a sensible way to pilot the tool before scaling it. York and Peel have run comparable systems for a period already without the kind of high-profile abuse scandal that would justify shutting the door on the technology entirely.
The Gap the Board Is Filling Itself
The problem is who is setting the rules. In May 2022, the Office of the Privacy Commissioner of Canada told Parliament that police use of facial recognition should be governed by clear statutory limits: restriction to "compelling purposes" such as serious crimes, an explicit ban on using FR to monitor peaceful protests, mandatory independent pre-authorization and audits, strict retention limits, and a prohibition on retaining data for people never charged (OPC, Recommended Legal Framework). That recommendation followed the OPC's 2021 finding that Clearview AI's scraping of billions of facial images constituted unlawful "mass surveillance" of Canadians — a case that showed regulators can punish misuse after the fact but cannot preempt it, because no statute defines what lawful police use of facial recognition actually looks like (OPC, Clearview AI news release).
More than four years after that recommendation, no such law exists. Bill C-27, which would have created Canada's first comprehensive AI governance statute, died on the order paper when Parliament prorogued, and nothing has replaced it. Into that vacuum steps a municipal police services board — an appropriate body to approve a procurement contract, but not one built to weigh Charter-level privacy tradeoffs for a technology that other Ontario forces are simultaneously adopting under their own, potentially divergent, policies.
Patchwork Governance Has Real Costs
This is not an argument that Ottawa's board acted improperly within the rules as they exist. It's an argument that the rules themselves are missing. Four Ontario forces are now running variations of the same Idemia-backed system under four separate board-level policies, with no single national oversight body auditing retention practices, no statutory floor on what counts as a permissible search, and no consistent standard a defence lawyer, journalist, or civil liberties group can point to across jurisdictions. The OPC's guidance documents are exactly that — guidance. They carry moral and reputational weight, but a police board that ignores them faces no binding consequence unless and until a Clearview-style investigation finds a violation years after deployment.
That asymmetry is bad for public trust and, contrary to the instinct that regulation is anti-innovation, bad for the technology's legitimate deployment too. Vendors like Idemia and the police services buying their tools would benefit from one clear, national standard rather than a patchwork of municipal board policies that could be challenged, reversed, or made politically radioactive by the next controversy. A proportionate federal statute — codifying the OPC's 2022 framework of necessity, proportionality, human review, and hard retention limits — would give both police and the public what neither currently has: predictability. Ottawa's board did what a responsible board can do with the tools available. Parliament still hasn't given it better ones.
What to Watch
- Whether Ontario's four-force shared network expands to a fifth city before any federal FR-specific bill is even tabled.
- Whether the successor to the AI and Data Act, if reintroduced, addresses law-enforcement facial recognition directly rather than deferring to sector guidance.
- Whether OPS publishes the transparency reporting the OPC's 2022 framework calls for once Phase 1 moves toward live use.