Nothing in Canadian employment law says a human must personally approve every rejection. What the law actually does is make sure the organization can't point at the software when asked to account for the decision.
Short answer
No general Canadian law prohibits a system from generating the rejection unaided. What Ontario's Employment Standards Act requires instead is disclosure — employers with 25 or more employees must state in the job posting that AI is being used to screen, assess or select applicants. Québec's Law 25 adds a different protection: where a decision is based exclusively on automated processing, the candidate can ask a person to review it.
Since January 1, 2026, a publicly advertised job posting in Ontario must include “a statement disclosing the employer's use, if any, of artificial intelligence to screen, assess or select applicants for the position” — but only for employers that count 25 or more employees on the day the posting goes up. The rule is a transparency duty, not a ban and not a consent requirement: it says nothing about whether a human has to review the output before an applicant is rejected.
Canada's federal, provincial and territorial privacy commissioners have said plainly that organizations using a generative AI system should “know that accountability for decisions rests with the organization, and not with any kind of automated system used to support the decision-making process.” An employer can't answer a rejected applicant, a tribunal, or a regulator by saying the software decided.
Québec's Law 25 doesn't require disclosure the way Ontario does — it creates a review right, and only where the decision is fully automated. The CAI's own guidance on hiring is specific: use of AI in a hiring process “devrait être communiqué aux candidats dès le début du processus” (should be communicated to candidates from the very start of the process), and where a rejection is “une décision fondée exclusivement sur un traitement automatisé” (a decision based exclusively on automated processing), the employer must tell the candidate and give them the chance to make representations to a staff member able to review it. Don't merge the two provinces' rules — Ontario discloses use, Québec triggers a review right, and each applies on its own terms.
The closest thing Canada has to a mandatory human-review rule for automated decisions is the federal Treasury Board Directive on Automated Decision-Making, which requires a pre-deployment impact assessment and a “meaningful explanation” after any automated administrative decision. It binds federal government departments, not private employers — there is no equivalent duty on an ordinary Canadian business hiring staff.
See also: exactly which employers the Ontario disclosure duty covers, how the ESA treats a decision made by a system generally, how this accountability principle plays out beyond hiring.
Evaluating what a hiring or screening tool actually does before it touches a real applicant pool is diligence work — ai-due-diligence covers it.
