A recent investigation and subsequent class-action lawsuit have brought to light a significant concern regarding the use of artificial intelligence in Ontario's correctional facilities. An AI-powered system designed to determine security classifications for incarcerated individuals is reportedly leading to a disproportionate placement of Black prisoners in maximum-security environments.
This tool, known as the Security Assessment for Evaluating Risk (SAFER), has been implemented in Ontario jails since 2021. Its function is to calculate a risk score for prisoners, drawing upon various factors such as arrest records and disciplinary history. However, an analysis of government data by a University of Toronto criminologist indicated that while Black individuals constitute approximately 5 percent of Ontario's general population, they represented nearly 27 percent of those held in maximum security between 2022 and 2025. Conversely, white individuals were notably underrepresented in this category, relative to their demographic proportion.
The legal action initiated by the law firm Koskie Minsky asserts that the province was allegedly aware of the potential for this AI tool to unfairly impact Black prisoners due to its reliance on historical justice system data, which itself reflects existing racial disparities in policing and charges. Furthermore, it is argued that while mitigation strategies were put in place for Indigenous prisoners, similar protective measures were not extended to Black prisoners. Internal documents from the ministry, referenced in the investigation, acknowledged that such assessment tools were likely to contribute to the overrepresentation of both Indigenous and racialized individuals in maximum security settings. As of now, neither the Ministry of the Solicitor General nor the purported designer of the SAFER tool has commented on the accuracy or methodology of the program, despite an ongoing review by the Ontario Ombudsman spurred by complaints regarding its discriminatory impact.
The findings underscore the urgent need for a critical examination of how algorithmic tools are integrated into sensitive areas like criminal justice. Ensuring equity and fairness in such systems is paramount to upholding justice and preventing the perpetuation of historical biases through technological means. This situation serves as a vital reminder that while AI offers powerful capabilities, its development and deployment must be accompanied by rigorous ethical oversight and a commitment to addressing systemic inequalities.