The Ottawa Police Service Board has approved the first phase of a facial recognition program, moving the force closer to using artificial intelligence to generate investigative leads. The decision, made at a June 22 meeting, allows the Ottawa Police Service (OPS) to lay the groundwork for the technology, including joining a database already used by police in York, Peel, and Halton regions.
Phase 1 Focuses on Implementation Readiness
According to a report presented to the board, Phase 1 concentrates on “implementation readiness” and does not yet authorize active use of facial recognition. The initial work includes analyzing safeguards, equity considerations, completing a privacy impact assessment, and developing a formal policy and set of procedures. These procedures will outline when the technology can be used, who can use it, and other ethical considerations.
The report specifies that the OPS will only use “legally obtained booking images” and will exclude “open-source or publicly scraped images.” This restriction aims to address some privacy concerns, but critics remain wary.
Privacy Concerns Raised by Experts
Gideon Christian, the University Research Chair of AI and Law at the University of Calgary, expressed concern about the technology. “The use of the technology is concerning because it turns body cameras — tools designed to watch and give accountability to the police — into tools the police use to watch the public,” he said.
Christian’s remarks highlight the tension between the accountability purpose of body-worn cameras and their potential use for surveillance. The OPS is also expanding its body-worn camera program under a $27.2-million contract with Axon, an American law enforcement technology developer, running until 2031.
AI Features in Body-Worn Cameras
Julie Kavanagh of the OPS’s media relations confirmed in an email that the police have procured Axon’s “AI Era Plan package, which includes access to several AI-enabled capabilities within the Axon ecosystem.” Two features the OPS will begin using are draft police reports and auto-translation. Kavanagh said auto-translation is an AI-assisted tool that supports real-time translation of audio and text across more than 50 languages.
The report drafting tool, called Draft One, “can generate an initial narrative based on body-worn camera footage and associated script,” Kavanagh added. This tool is designed to streamline police reporting but raises questions about accuracy and oversight.
Next Steps and Broader Implications
The OPS’s move toward facial recognition comes as police forces across Canada explore AI technologies. The network the OPS hopes to join compares human faces from digital images or video frames against existing databases, according to the Halton Regional Police Service. While the board has approved Phase 1, active use of facial recognition still requires further approval and completion of the outlined assessments.
The contract with Axon includes access to additional AI capabilities that may be evaluated in the future, Kavanagh noted. As the OPS proceeds, the balance between investigative efficiency and civil liberties will remain a key point of debate.
