Police departments across Ontario are taking matters into their own hands by establishing AI governance policies, stepping into a regulatory gap where provincial and federal governments have yet to act. London Police Services Board approved new rules around artificial intelligence use at its meeting this week, joining York, Peel, and Toronto police services in creating local frameworks that prioritize human oversight, privacy protection, and public trust.

Why Are Police Forces Creating Their Own AI Rules?

The absence of a provincial AI framework has left police services boards responsible for establishing governance expectations around emerging technologies. As artificial intelligence becomes increasingly embedded in law enforcement operations, from facial recognition to predictive policing tools, departments recognize they cannot wait for government guidance . The London Police Services Board's decision reflects a broader understanding that AI technologies offer operational efficiency gains but introduce significant risks that demand immediate attention.

"AI technologies are becoming increasingly embedded in policing. While they offer opportunities for efficiency, they also introduce risks related to privacy, bias and public confidence," said Ryan Guass, board chair at London Police Services Board.

Ryan Guass, Board Chair, London Police Services Board

This decentralized approach to AI governance reveals an important reality: law enforcement agencies cannot afford to wait for top-down regulation. Instead, they are learning from each other's experiences and adapting best practices developed by larger urban police services.

What Does London's AI Policy Actually Require?

London Police's framework establishes several concrete requirements for how the department can deploy AI technologies . The policy mandates that artificial intelligence systems remain subject to meaningful human oversight, ensuring that no critical law enforcement decision relies solely on algorithmic recommendations. Additionally, the framework requires that any use of AI be justified, proportionate, and consistent with legal and ethical standards.

The policy also establishes compliance mechanisms to ensure ongoing accountability. An "AI Technology Compliance and Risk Report" will be presented to the police board annually, creating a structured process for reviewing how AI systems perform and whether they introduce unintended harms . Portions of that report may engage operational, legal, or security sensitivities, so a public-facing summary will likely be required, ensuring transparency without compromising sensitive investigations.

The framework explicitly ties AI use to Canada's legal landscape, requiring that all AI technologies comply with applicable laws, including the Canadian Charter of Rights and Freedoms, human rights legislation, privacy laws, and policing legislation . This legal grounding ensures that efficiency gains do not come at the expense of constitutional protections.

How to Implement Responsible AI Governance in Law Enforcement

• Establish Human Oversight Requirements: Mandate that AI systems support human decision-making rather than replace it, ensuring officers retain authority over critical enforcement actions and that algorithms serve as tools for analysis rather than autonomous decision-makers.
• Create Annual Compliance Reporting: Develop structured reporting mechanisms that assess AI system performance, identify bias or fairness issues, and track whether deployed technologies actually deliver promised benefits without introducing new risks to public trust.
• Align Technology Use With Legal Standards: Ensure all AI deployments comply with constitutional protections, human rights legislation, and privacy laws specific to your jurisdiction, rather than adopting tools that may work in other regions but violate local legal obligations.
• Conduct Proportionality Assessments: Before deploying any AI system, evaluate whether the operational benefits justify the identified risks to privacy, bias, and public confidence, documenting this analysis for accountability purposes.
• Develop Public-Facing Transparency Mechanisms: Create summaries of compliance reports that inform the public about how AI is being used in policing while protecting operational security and ongoing investigations.

The London Police approach reflects a pragmatic recognition that police boards cannot simply adopt AI tools without establishing clear governance structures. By drawing on measures taken by police services boards in York, Peel, and Toronto, London Police demonstrates how smaller departments can benefit from the groundwork laid by larger urban forces .

What Does This Mean for the Broader AI Governance Conversation?

The emergence of police-led AI governance frameworks highlights a critical gap in government regulation. Rather than waiting for provincial or federal authorities to establish comprehensive AI policy, law enforcement agencies are becoming de facto regulators of their own technology use. This bottom-up approach has advantages: it allows departments to tailor policies to their specific operational contexts and community needs. However, it also creates inconsistency, as different police services may adopt different standards for the same technologies.

London's policy requirement that "AI must remain subject to meaningful human oversight, and that its use must be justified, proportionate, and consistent with legal and ethical standards" establishes a baseline that other departments can reference . This language echoes principles emerging in broader AI governance discussions, suggesting that even without formal government regulation, law enforcement is converging on shared expectations about responsible AI deployment.

The annual compliance reporting requirement is particularly significant. By creating a structured process for reviewing AI system performance, police boards are establishing accountability mechanisms that government regulators might eventually adopt as baseline standards. This could position early-adopting police services as models for how other sectors should govern emerging technologies.

As AI continues to reshape law enforcement operations, the London Police decision signals that departments cannot rely on government to move quickly enough. Instead, they are taking responsibility for ensuring that efficiency gains do not undermine public trust, privacy protections, or constitutional rights. Whether this decentralized approach ultimately leads to more effective AI governance or creates problematic inconsistencies across jurisdictions remains an open question, but it demonstrates that the pressure to govern AI responsibly is coming from within institutions themselves, not just from external advocates or regulators.