AI Screening Tools and Canadian Law: What Every SMB Hiring Manager Needs to Know in 2026

If your business has 25 or more employees in Ontario and you posted a job in the last six months without disclosing that your applicant tracking software uses AI to rank or filter candidates, you are out of compliance with the Employment Standards Act. You may not have known the rule existed.

The requirement came into force on January 1, 2026, under Ontario's *Working for Workers Four Act*. It is Canada's first provincial law explicitly regulating how employers use artificial intelligence in hiring — and it arrived with a corporate penalty of up to $100,000 for violations.

The rule itself is narrow: it requires disclosure, not audits, not testing for bias, not any mechanism for a rejected candidate to challenge an algorithmic outcome. What it does do is establish, formally and legally, that employers are responsible for knowing whether AI touches their hiring process and for telling applicants when it does. Most Canadian SMBs have not done that analysis.

The Tools Most Canadian SMBs Are Already Using

Hiring AI is not a hypothetical technology. It is embedded in the tools most employers have been using for years without thinking of them as "artificial intelligence."

Resume parsing and ranking — the core function of almost every modern applicant tracking system (ATS) — is AI. Tools like Workable, Greenhouse, BambooHR, and Lever all use algorithmic matching to score resumes against job descriptions and surface the highest-ranked candidates first. Scheduling assistants that identify time slots and send automated follow-ups are AI. Chatbots that pre-screen applicants through structured question sequences are AI. Video interview analysis platforms that score candidates on response content, pacing, or affect are AI.

SHRM's State of AI in HR 2026 found that recruiting is the most AI-intensive function across all HR disciplines — 39% of HR professionals who have adopted AI use it in recruiting, making it the single most common application. A February 2024 Gartner survey of 179 HR leaders found that 38% were piloting or had already implemented generative AI in HR — double the proportion from just six months earlier.

The BDC has found that 27% of Canadian entrepreneurs use AI tools without realizing it. The same pattern almost certainly applies to hiring: many Canadian SMBs are using AI-powered ATS features they installed years ago, often as default settings, without having consciously decided to deploy "AI" in their recruitment process.

The AI in the tool does not know it needs to be disclosed. That responsibility now rests with you.

What Ontario's Law Actually Requires

Ontario's *Working for Workers Four Act* amended the *Employment Standards Act, 2000* to add a new disclosure obligation for job postings. As of January 1, 2026:

• Who must comply: Employers with 25 or more employees on the day a publicly advertised job posting is posted
• What must be disclosed: A statement that the employer uses AI to "screen, assess, or select" applicants — if it does
• Where the disclosure appears: In the publicly advertised job posting itself
• Level of detail required: Simple. The Ontario Ministry of Labour has confirmed that "it is not necessary to provide a detailed description of the artificial intelligence system." A one-line statement that AI is used in the screening process is sufficient.
• Penalty for non-compliance: Up to $100,000 for corporations under the ESA enforcement regime, higher for repeat violations

The law has generated significant uncertainty among Ontario employers because the Ministry has not yet published guidance on what counts as "artificial intelligence." Legal consensus is that any algorithmic tool that scores, ranks, or filters candidates likely qualifies — covering most ATS systems with any automated screening features.

What the law notably does *not* require: a pre-deployment equity audit, testing for disparate impact under the *Human Rights Code*, or any appeal mechanism for a rejected applicant. Ontario's disclosure rule is a transparency measure, not a fairness one. The fairness obligations come from other law.

Beyond Ontario: The Laws That Already Apply Everywhere

Ontario's ESA amendment is the most concrete compliance deadline for SMBs right now. But it does not operate in isolation, and it is not the only law that governs AI in Canadian hiring.

Quebec Law 25

Quebec's *Act Respecting the Protection of Personal Information in the Private Sector* — fully in force since September 2023 — creates explicit rights around automated decision-making. Where an organization makes a decision about an individual based exclusively on automated processing, it must:

• Inform the individual that an automated system was used
• Provide the individual with the right to have a human review the decision
• Disclose the type of personal information used and the main factors that influenced the outcome

If your hiring process uses AI to filter out applicants before any human sees the file, Quebec Law 25's automated decision provisions apply — and they extend to job applicants, not just customers. Any Quebec-based SMB using an ATS that makes first-pass screening decisions without human review is exposed under provincial privacy law today.

PIPEDA and Job Applicant Data

For federally regulated employers — banks, airlines, telecommunications companies, and other federal undertakings — PIPEDA governs the collection and use of job applicant personal information. The OPC's guidance on Privacy in the Workplace establishes that applicant personal data must be collected only for the purpose of evaluating suitability for the role, disclosed as being used in an AI screening process, and retained only as long as necessary after a hiring decision — typically three to six months.

The OPC's December 2023 Principles for Responsible Generative AI add a critical accountability principle: responsibility for AI-influenced decisions rests with the organization, not the AI system or its vendor. If a third-party ATS discriminates or mishandles applicant data, the employer is accountable under Canadian privacy law. Contractual indemnification from the vendor does not transfer that accountability away.

The Canadian Human Rights Act

The Canadian Human Rights Act prohibits employment discrimination on 13 grounds including race, sex, age, disability, national or ethnic origin, and sexual orientation. Provincial human rights codes extend equivalent protections to all provincially regulated employers.

AI hiring tools create a specific human rights exposure: systems trained on historical hiring data replicate past biases. A tool trained on a decade of decisions in a male-dominated industry learns to prefer profiles that resemble past hires — which is a proxy for gender. A tool that uses postal codes, employment gaps, or educational institution names as scoring features will proxy-discriminate on race, disability, and socioeconomic status.

The Canadian Human Rights Commission's 2024 Annual Report to Parliament states directly: "Bias and discrimination amplified by artificial intelligence are real and complex, but can be easily overlooked, and left unchecked, can cause deep and long-standing harm to individuals, communities and organizations." The CHRC collaborated with the Ontario Human Rights Commission and the Law Commission of Ontario to release a Human Rights AI Impact Assessment Tool in fall 2024 specifically for evaluating the discrimination risk in automated systems — including hiring tools.

Discriminatory AI outputs are legally actionable today, without waiting for new AI legislation. The OECD Employment Outlook states that AI can either improve hiring fairness or reinforce systematic bias, "particularly where [AI systems are] used in highly impactful contexts such as employment" — depending entirely on how the system is designed and what data it was trained on. Employers cannot assume fairness by default.

The Federal Regulatory Picture

Canada's proposed *Artificial Intelligence and Data Act* (AIDA) — contained within Bill C-27 — died on the Order Paper when Parliament was prorogued in January 2025. The bill would have explicitly classified hiring, recruitment, and promotion AI as "high-impact AI systems" requiring mandatory risk assessments before deployment. It never became law.

Prime Minister Mark Carney's government launched the "AI for All" national AI strategy on June 4, 2026, with commitments to introduce new AI legislation over the next five years. The current federal approach remains voluntary — relying on ISED's Voluntary Code of Conduct on the Responsible Development and Management of Advanced Generative AI Systems rather than binding rules.

The practical implication for Canadian SMBs: no federal AI law governs hiring tools today. But PIPEDA, the CHRA, and provincial human rights codes already apply — and the direction of federal AI legislation is unmistakable. The AIDA Companion Document explicitly named hiring and recruitment as applications requiring the most stringent protections. When federal AI legislation returns to Parliament, SMBs using AI in hiring should expect to be in scope.

A Practical Checklist for Canadian SMBs

The legal environment is still forming, but the current obligations are real and the direction of travel is clear. Here is a starting framework.

Audit your hiring tools

Map every tool that touches your recruitment workflow and determine whether it uses AI to score, rank, or filter candidates at any stage. Check vendor documentation and release notes. If a tool has a "match score," "candidate ranking," or "smart shortlist" feature, it is almost certainly AI under Ontario's definition.

Update your Ontario job postings immediately

If you have 25 or more employees and publicly advertise positions in Ontario, add a disclosure statement to all active and future postings. A single line — "We use artificial intelligence tools to assist in the screening of applications" — meets the Ministry of Labour's stated standard. Review any job posting template or ATS default that auto-generates your postings.

Keep a human in the loop for all screening decisions

Never let AI make the final determination on whether a candidate advances or is rejected without documented human review. This is the single most effective protection against human rights liability, meets Quebec Law 25's human review requirement, and aligns with the compliance expectation that federal AI legislation will almost certainly impose when it arrives.

Apply data minimization to applicant records

Collect only what you need for the hiring decision — work history, qualifications, and references. Set data retention limits: three to six months after a hiring decision is standard practice. Do not collect Social Insurance Numbers during the application process (they are only needed post-hire, for payroll purposes). Confirm where your ATS stores applicant data; most major platforms offer Canadian data residency options for tenants configured on their Canadian region.

Ask your vendor the right questions

Before signing or renewing any AI recruitment tool contract, ask: What training data was used? Has the system been independently audited for disparate impact on protected grounds? What bias controls are in place? What data does the system use as features, and could any of those features proxy-discriminate? Reputable vendors will have documentation. Vendors who cannot answer the questions are telling you something important about the risk you are taking on.

Sources

• Statistics Canada. *Analysis on Artificial Intelligence Use by Businesses in Canada, Second Quarter of 2025.* statcan.gc.ca
• SHRM. *State of AI in HR 2026.* shrm.org
• Gartner. *Gartner Finds 38% of HR Leaders Are Already Piloting or Implementing Generative AI.* February 2024. gartner.com
• BDC. *New BDC Study Reveals 27% of Canadian Entrepreneurs Don't Know They're Using Artificial Intelligence.* bdc.ca
• Osler, Hoskin & Harcourt LLP. *AI in Hiring: Ontario Employers Grappling with New Job Posting Disclosure Requirement.* osler.com
• Fasken. *New Rules for Automated Decision Making Under Quebec Bill 64 (Law 25).* fasken.com
• Office of the Privacy Commissioner of Canada. *Privacy in the Workplace.* priv.gc.ca
• Office of the Privacy Commissioner of Canada. *Principles for Responsible, Trustworthy and Privacy-Protective Generative AI Technologies.* December 2023. priv.gc.ca
• Canadian Human Rights Commission. *2024 Annual Report to Parliament.* chrc-ccdp.gc.ca
• Ontario Human Rights Commission. *Human Rights AI Impact Assessment Tool.* ohrc.on.ca
• OECD. *Employment Outlook 2023: Artificial Intelligence, Job Quality and Inclusiveness.* oecd.org
• Prime Minister of Canada. *Prime Minister Carney Launches AI for All: Canada's New National Artificial Intelligence Strategy.* June 4, 2026. pm.gc.ca
• Innovation, Science and Economic Development Canada. *The Artificial Intelligence and Data Act (AIDA) — Companion Document.* ised-isde.canada.ca
• BDC. *Labour Shortage: Addressing the Talent Gap.* bdc.ca
• Gartner. *Gartner Survey: 88 Percent of HR Leaders Say Their Organizations Have Not Realized Significant Business Value from AI Tools.* October 2025. gartner.com

Not sure whether your hiring tools trigger Ontario's new disclosure requirement — or what your obligations are under PIPEDA and provincial human rights law? Our AI Advisory team helps Canadian SMBs conduct AI readiness assessments that include your hiring stack, privacy obligations, and compliance exposure. Book a free consultation to get a clear picture before your next job posting.