IEEE Technology and Society Magazine - September 2022 - 50

Technological Stewardship and Responsible Innovation
Risk Assessment Tools
on Trial: AI Systems Go?
Neha Chugh
Concordia University
Montreal, QC H3G 1M8, Canada
 In the 2015 Canadian case of Ewert v. Canada,
risk assessment tools were put on trial in Canada's
Federal Court and eventually at the Supreme Court
of Canada, their efficacy was challenged, and their
reliability was upended [1], [2]. Risk assessment
tools are used by the justice system to present a
calculated prediction of an offender's risk of future
criminal behavior. These tools are viewed by the
legal community as precursors to machine learning
and technological advancements in artificial intelligence
(AI) in the criminal justice system, using algorithmic
and data-driven decision-making to provide
courts with predictions regarding an offender's risk
if released.
Data-driven decision-making is the process of
forecasting, predicting, or arriving at a recommendation
using preexisting datasets. Risk assessment tools
used by correctional facilities are a type of datadriven
decision-making. Data-driven decision-making
can also be algorithmic decision-making, an
algorithm being the procedure or formula used to
process the data. AI is the most advanced form of
algorithmic decision-making in the justice system,
using machine learning and computational methods
to adapt and process information.
In the Canadian context, AI is reforming rapidly:
machine learning and algorithmic decision-making
are beginning to proliferate in the justice system. The
common law is developing slowly in response to the
findings in Ewert, despite red flags being raised by
scholars regarding the frailties of data-driven decision-making
in the criminal justice system. Canadian
law has traditionally prioritized individual and
community factors, whereas algorithms informing
risk assessment tools and artificial intelligence are
focused on averages, patterns, and calculations.
The starting point of assessing our readiness to
embrace artificial intelligence in the Canadian criminal
justice system emerges from Ewert and then
moves into a discussion of broader, systemic implications
as well as input from international experiences.
The following lessons have been realized for future
implications of AI in the criminal justice system.
1. Risk assessment tools have been in use by the justice
system for decades, despite concerns being
flagged within justice system participants regarding
their unreliability, inaccuracy, and limited predictive
value.
2. Our courts are tasked with consideration of individualized
factors unique to colonial history:
residential schools, racism, substance abuse,
loss of land and dislocation, separation from parents
and families of origin, and physical/sexual/
spiritual abuse. The courts have identified a crisis
faced by Canada's Aboriginal people in our justice
system, marred by systemic discrimination
and racism [16].
3. Criminal justice systems in Canada and jurisdictions
around the world continue to rely on risk
assessment tools. Eventually, will data-driven decision-making
replace our existing tools through
AI, reproducing or perpetuating the biases that
already exist within crime data [6, p. 84]?
Digital Object Identifier 10.1109/MTS.2022.3197123
Date of current version: 16 September 2022.
50
0278-0097/22©2022 IEEE
Ewert v. Canada: On trial, Jeffrey Ewert, an Aboriginal
person, is serving a life sentence in Canada's
federal penitentiary system for murders committed in
1984 [2]. Ewert's eligibility for parole was adversely
affected by the Correctional Services of Canada's
(CSC) representations of Ewert's results from the Hare
IEEE Technology and Society Magazine
IEEE Technology and Society Magazine - September 2022

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - September 2022
