Public Service AI Decisions to Face New Transparency Rules

The federal government plans to impose new transparency rules on automated decision-making systems used by the public service and federal agencies, a policy document released Wednesday shows.

The rules would require agencies to publish descriptions of the AI tools they deploy, disclose when an automated system makes or supports a decision touching a person's rights, and offer an avenue for human review. The directive covers systems that assess, rank, recommend, or deny benefits, permits, or immigration outcomes.

Treasury Board President Anita Anand, who oversees the government's digital strategy, said the policy aims to make algorithmic decisions "visible, explainable, and contestable" for Canadians affected by them.