Summaries, not legal advice. We summarize what each document says. Read the original, and ask your service's legal counsel, before relying on it.
The legal gap
The federal, provincial and territorial privacy commissioners said in 2022 that the current legislative context for police use of facial recognition is insufficient. [Joint guidance, May 2022] The Toronto Police Service Board policy states: "No current statutes or regulations fully govern the use of AI technologies in Ontario or Canada." [TPSB policy]
Clearview AI: what regulators found
- October 2019The RCMP purchased two Clearview AI licences, and members also used free trial accounts. [OPC, 2021]
- TorontoToronto Police Service officers used Clearview AI in 84 criminal investigations over about three and a half months, before the chief found out and ordered them to stop, according to CBC (Dec 2021). [CBC News]
- 21 February 2020The OPC and the privacy authorities of Quebec, Alberta and British Columbia opened a joint investigation of Clearview. [OPC, 2021]
- February 2021The four authorities found Clearview's practices were mass surveillance and contravened PIPEDA and provincial privacy law. [Joint news release, Feb 2021] [Joint findings, 2021]
- June 2021The OPC found the RCMP contravened the Privacy Act. [OPC news release, June 2021]
- May 2022Final joint guidance for police on facial recognition was published. [Joint guidance, May 2022]
On the RCMP finding, the OPC explained that a government institution cannot collect personal information from a third party that collected it unlawfully. It also said the RCMP first erroneously told the OPC it was not using Clearview AI, and did not satisfactorily account for most of its searches. [OPC, 2021]
"police should not use FRT just because it is thought to be 'useful' for law enforcement in general. Police should have a specific reason to use the technology and it should be based on evidence."
2022 joint guidance on facial recognition
The guidance is issued jointly by the privacy authorities of every province and territory and the federal OPC. It sets out what police agencies' privacy obligations are under current law. Its headings include lawful authority, the Charter, necessity and proportionality, privacy impact assessments, monitoring and re-assessment, accuracy, accountability, procurement, training, data minimization, retention, testing, data security, and transparency. [Joint guidance, May 2022]
"This guidance should not be read as justifying, endorsing or approving the use of FR by police agencies."
Toronto Police Service Board policy
The Board approved its Use of Artificial Intelligence Technology policy on 28 February 2022 (amended 11 January 2024). It covers new or enhanced AI technology, and previously approved AI used for a new purpose. [TPSB policy] The guiding principles are legality, fairness, reliability, justifiability, personal accountability, organizational accountability, transparency, privacy and meaningful engagement. [TPSB policy]
Two features stand out. Where an AI-assisted decision may lead to charges, the possible influence of the AI must be included in the disclosure provided to the Crown. And the Board says reviews should be conducted in public to the greatest degree possible. [TPSB policy] The Ontario Human Rights Commission noted that the service's AI inventory identified one facial recognition technology as "high risk". [OHRC on TPSB policy]
CACP draft principles
The Canadian Association of Chiefs of Police has developed ethical principles for responsible AI use in Canadian police services. The principles are grouped under fairness; data ethics; governance, impartiality and auditability; rights and self-determination; and control, security and reliability. They align with the CACP and CENTRIC programme Canadian AI Accountability for Policing and Security (C-AIPAS). [CACP principles (FR), July 2026] [CACP Data & AI Summit, Oct 2026]
In October 2026 the CACP asked members for feedback on the proposed principles, describing them as guidance rather than imposed rules. [CACP news alert, Oct 2026] The text we read is the French version, dated 20 July 2026. Our unofficial translation of its opening position: for Canadian police, the strategic question is no longer whether AI will be used, but whether its use will be structured to preserve public trust and the rule of law. [CACP principles (FR), July 2026]
Charter and human rights positions
These are positions taken by research and oversight bodies, not court rulings.
- Citizen Lab (2020) argues that algorithmic policing can engage Charter rights, including privacy, expression, equality, liberty and due process. [Citizen Lab, 2020]
- The Canadian Human Rights Commission (2022) said police use of facial recognition should be prohibited through a moratorium until a new legal and policy framework is in place. [CHRC, 2022]
Legislation
As reported by The Canadian Press in April 2025, the federal Artificial Intelligence and Data Act died when Parliament was prorogued on 6 January 2025, and the Canadian Civil Liberties Association noted it applied only to the private sector. The same report says Ontario was the first province to adopt legislation about government agencies using AI, and that the Law Commission of Ontario said the bill leaves unclear how far it covers police. [Blue Line / Canadian Press, April 2025] We have not checked later legislative developments; see the corrections policy and tell us if this is out of date.
A comparison: the UK covenant
In the UK, all members of the National Police Chiefs' Council endorsed a Covenant for Using AI in Policing on 28 September 2023. It says developers and users of AI within policing must give due regard to its principles, and that AI in policing "can be contentious". [NPCC AI Covenant]