The LCO presented 29 recommendations to facilitate responsible AI use by Crowns and courts
The Law Commission of Ontario has recommended the drafting of new provincial legislation and the establishment of governance and risk management frameworks for artificial intelligence systems used in the criminal justice system.
The LCO’s landmark final report on “AI in Criminal Justice” is the first in Canada to examine the uses, benefits, and risks of integrating AI systems into the nation’s criminal justice system. In it, the LCO presented 29 recommendations to facilitate responsible AI use by police, Crowns, courts, and other institutions.
The recommendations include developing a governance framework for AI systems utilized in criminal justice. This framework must set clear legal and policy guidance for governments, law enforcement agencies, courts, Crowns, defence counsel, and legal aid to support public safety, legal rights protection, and public trust building.
The LCO suggested that the framework should include key elements like the mandatory public disclosure of AI system use in criminal justice, a dedicated and required impact assessment for such systems, continuous auditing and evaluation, and guarantees for judicial independence. Moreover, the commission urged the Ontario government to either modify the Enhancing Digital Security and Trust Act, 2024, or introduce complementary laws to enshrine in law governance principles and mandates.
These laws must include a preamble subjecting AI systems to a precautionary principle. Systems must adhere to the Charter, principles of criminal procedural fairness, privacy law, human rights law, criminal common law, and the principle of judicial independence. The legislation should also set mandatory risk categories for AI systems as well as reporting requirements to the province’s Inspector General of Policing and Law Enforcement Complaints Agency, which will then publish public report examining AI use by law enforcement services every year.
The LCO also pushed for the establishment of a joint Federal-Provincial Criminal AI Evidence Rules Committee and consultation process to examine and recommend basic safeguards for AI evidence rules in relation to criminal matters. It called on Ontario’s Ministry of the Attorney General to boost funding and other resources for Legal Aid Ontario to support criminal defence counsel in responding to and defending cases where AI evidence is presented.
The commission recommended that Legal Aid Ontario set up an AI Centre of Excellence to provide training on criminal justice AI-related laws and procedure, set standards for professional competence in AI legal research and advocacy, conduct legal research on AI issues and cases, and offer support for individual and systemic AI test cases.
“Many AI technologies can improve public safety, improve police investigations and improve the efficiency and fairness of criminal proceedings. At the same time, criminal AI systems such as predictive policing, [facial recognition technology] and generative AI systems have proven risks to human rights, criminal procedure, privacy and access to justice. These risks affect racialized and low income communities most acutely,” report authors Ryan Fritsch and Nye Thomas wrote in the executive summary. “To date, no Canadian government has enacted laws or regulations governing the use of AI by Canadian police or criminal courts.”
Fritsch leads the LCO’s Criminal AI Project and is a counsel with the organization, while Thomas is LCO executive director.
