In a campaign to recruit executives, the Department of National Defence (DND) used products from Toronto-based Knockri and Waterloo, Ont.-based Plum, which claim to use AI to reduce bias. It did not do an algorithmic impact assessment (AIA), as required for most government AI projects, and the privacy commissioner’s office said it also did not submit a privacy impact assessment. (The Globe and Mail) 

Talking point: Ottawa has talked up its ethical use of AI, enacting a directive on automated decision-making in April 2019 designed to ensure its outcomes are “accurate, consistent, and interpretable” and minimize the negative ones. But so far only one AIA—the key requirement of the policy—has been filed. DND said it didn’t complete the questionnaire because the AI system didn’t make the final call on whom to hire. In November 2019, The Logic reported the government didn’t consider Palantir’s controversial work with U.S. law enforcement as part of an ethics test for companies seeking to pre-qualify to bid on government AI projects.