54% of Ontario Employers Use AI in Hiring. Almost None Disclose It.
Ontario now requires employers with 25+ staff to disclose AI use in job postings. Most have not complied. Here is what the law requires.
The specific disclosure obligations Ontario employers face under the Working for Workers Four Act, which AI hiring tools trigger the requirement, and a four-step compliance process to implement before your next job posting.
Ontario's AI hiring disclosure requirement mandates that employers with 25 or more employees disclose when artificial intelligence is used to screen, assess, or select job applicants in any publicly advertised job posting. The requirement took effect January 1, 2026 under the Working for Workers Four Act (Bill 149) and carries penalties up to $100,000 CAD per conviction under the Employment Standards Act.
54% of employers already use AI in hiring, but fewer than 1 in 10 disclose it in job postings. Penalties under the ESA reach $100,000 CAD per conviction (Ontario ESA).
More than half of Canadian hiring managers say their organizations use artificial intelligence in recruitment (Express Employment Professionals). Since January 1, 2026, Ontario employers with 25 or more employees who use AI to screen, assess, or select job applicants must disclose that use in every publicly advertised job posting and every associated application form (Osler). Almost none have done it.
The gap between usage and disclosure is a compliance failure that carries fines up to $100,000 per conviction under the Employment Standards Act (Canada Employment & Human Rights Law). Most employers have not complied because they do not know which of their tools count as "artificial intelligence" under the law.
What the Law Actually Says
The AI disclosure requirement comes from the Working for Workers Four Act, 2024, which received Royal Assent on March 21, 2024 (Ontario Legislature). It amends the Employment Standards Act to add a new obligation: any employer with 25 or more employees on the day a job posting goes live must state whether AI is used to screen, assess, or select applicants for that posting.
The definition of "artificial intelligence" in the legislation is broad. It covers "a machine-based system that, for explicit or implicit objectives, infers from the input it receives in order to generate outputs such as predictions, content, recommendations or decisions that can influence physical or virtual environments" (Littler).
That definition captures any system that infers, predicts, ranks, recommends, or filters based on learned patterns, extending well beyond chatbots and generative AI. For most employers, this means the tools already running inside their applicant tracking systems.
Which Tools Trigger the Disclosure
The Ministry of Labour has not issued detailed guidance on what qualifies (Osler). That silence is the core compliance problem. Employers are legally obligated to disclose AI use but have no framework for determining which tools fall within scope.
Based on the statutory definition and analysis from Osler, Hicks Morley, and SpringLaw, these categories of hiring tools likely trigger the obligation:
Resume screening software that filters, ranks, or scores applicants based on keyword matching, experience patterns, or predictive models. Most enterprise applicant tracking systems (Workday, Greenhouse, Lever, iCIMS, SmartRecruiters) include AI-powered screening features, often enabled by default.
When a platform evaluates facial expression, tone, speech patterns, or word choice to score candidates, it falls within the definition. HireVue and similar video interview analysis tools are in scope.
LinkedIn Recruiter's AI-powered candidate matching recommendations count. Any engine that recommends applicants to recruiters based on inferred fit scores triggers the obligation.
Algorithmic scoring that predicts job performance or personality traits also qualifies. Criteria Corp, Pymetrics (now Harver), and similar pre-employment assessment platforms are covered.
The same applies to chatbot-based screening. Paradox (Olivia), Phenom, and XOR ask qualifying questions and advance or eliminate candidates based on responses.
Tools marketed as "automation" rather than "AI" can still trigger the disclosure if they infer from input to generate predictions or recommendations about candidates (SpringLaw), because the legal test is functional, not branding.
An Ontario employer with 40 staff uses BambooHR for recruitment. The platform's default settings include AI-powered candidate ranking. The hiring manager does not know the feature is active. Under Bill 149, every posting made through that system requires AI disclosure, whether or not the hiring manager intentionally enabled the screening. The compliance gap is about tool awareness, not intent.
Why Most Employers Have Not Complied
Three factors explain the gap between the law taking effect and actual compliance.
The Ministry of Labour has not published guidance (Osler). Employers are expected to comply with a disclosure obligation when the regulator has not clarified what "screen," "assess," and "select" mean in practice. An employer using Workday's automated resume sorting does not necessarily know whether Workday considers that feature "artificial intelligence."
Many employers do not realize their existing tools use AI. The Express Employment Professionals survey found that while 54% of Canadian hiring managers say their organizations use AI, 21% use it regularly, up from 16% the previous year (Express Employment Professionals). The remaining 33% are using AI intermittently through features embedded in platforms they already pay for. Features like "recommended candidates" or "smart screening" qualify but rarely get flagged internally as AI usage requiring legal disclosure.
Enforcement has not been aggressive yet. Applicants can file complaints under the Employment Standards Act (Hicks Morley). Regulators have not prioritized monetary enforcement for these new requirements. The regulatory pattern with employment standards is slow rollout followed by targeted enforcement after a complaint wave establishes precedent.
Not sure where AI fits in your operations?
Take the Free AI Readiness Assessment →The Human Rights Dimension
The disclosure requirement exists for a reason beyond transparency. AI tools trained on historical hiring data replicate and amplify existing patterns of discrimination. The Ontario Human Rights Commission has flagged AI use in employment as a growing risk, citing the potential for indirect discrimination through algorithmic bias (OHRC).
A review of the Peel District School Board found that an algorithmic technology used to vet prospective teaching candidates had inappropriately screened out racialized candidates who were otherwise qualified (HRD Canada).
In November 2024, the OHRC published its Human Rights AI Impact Assessment, a practical framework for identifying and mitigating human rights risks across the AI lifecycle (OHRC). Employers using AI in hiring face not only ESA disclosure obligations but also potential human rights liability under the Ontario Human Rights Code if their tools produce discriminatory outcomes, disclosed or not.
How Ontario Compares to Other Jurisdictions
Ontario's approach is lighter than what New York City has required since 2023. NYC Local Law 144 prohibits employers from using automated employment decision tools unless the tool has undergone an independent bias audit within the preceding year and results are publicly posted. Violations carry penalties of $500 to $1,500 per day (NYC DCWP).
Ontario requires disclosure but not auditing. New York requires auditing but applies only to automated decision tools, not the broader "artificial intelligence" definition Ontario uses (Data Protection Report).
The pattern across jurisdictions: disclosure first, then audit mandates, then enforcement escalation. The EU AI Act classifies recruitment AI as "high-risk," requiring conformity assessments before deployment (Osler). Canada's federal Artificial Intelligence and Data Act is pending. Ontario employers who wait for audit mandates before conducting one will find themselves behind when requirements escalate.
For Ontario public sector institutions, regulation is already ahead of the private sector: Bill 194 created mandatory cybersecurity and AI accountability requirements with IPC breach reporting effective March 2026.
What Compliant Disclosure Looks Like
Without Ministry guidance, employers should err on the side of transparency. Based on legal analysis from SpringLaw and Hicks Morley, compliant job postings should include:
A clear statement that AI is used in the hiring process. Example: "This organization uses artificial intelligence tools to assist with screening and assessing applicants for this position."
Specificity about which stage uses AI. Example: "AI is used for initial resume screening. All shortlisted candidates are reviewed by human recruiters before interview decisions are made."
The same disclosure on any associated application form where candidates submit materials (Littler).
Employers must also retain copies of every public job posting and associated application form for three years after removal (HRPA). Your HR team needs an archiving system, not just a disclosure template.
Employers who conduct an HR technology audit and add a standard disclosure statement to their job posting template can achieve compliance in under a week. The four-step process below prevents both current ESA exposure and positions the business ahead of broader AI governance requirements arriving at the federal level.
What to Do Now
Inventory every tool in your hiring process. From the job board where the posting appears, through the ATS that receives applications, to the assessment platform that scores candidates. Map the full pipeline.
For each tool, determine whether it uses inference, prediction, recommendation, ranking, scoring, or filtering based on learned patterns. Contact your vendors and ask for a written statement on whether their tools use AI within the Ontario ESA definition. Some vendors have begun issuing these statements proactively.
Update your job posting templates to include disclosure language. Apply it to every active posting retroactively.
Establish a three-year retention system for all postings and associated application forms.
Consider a voluntary bias audit of your AI hiring tools. Ontario does not require this yet, but the OHRC's Human Rights AI Impact Assessment provides the framework. Documenting that you tested for bias before a complaint is filed is the strongest defense available.
DeployLabs builds the systems that make compliance operational. Our $2,500 AI assessment maps your current technology stack, identifies which tools meet Ontario's AI definition, and produces the documentation your HR and legal teams need.
Book a discovery call to find out in 30 minutes whether your hiring stack triggers the obligation.
- Ontario employers with 25+ employees must disclose AI use in job postings since January 1, 2026 under the Working for Workers Four Act. Penalties reach $100,000 per conviction.
- Most non-compliance is accidental: AI screening features are embedded in HR platforms like Workday, BambooHR, and Greenhouse, often enabled by default.
- The Ministry of Labour has not issued guidance on which tools qualify, leaving employers to interpret a broad statutory definition of "artificial intelligence."
- Beyond ESA penalties, employers face human rights liability if AI tools produce discriminatory outcomes, whether disclosed or not.
- The regulatory pattern moves from disclosure to audit mandates. Employers who build governance frameworks now avoid scrambling when requirements escalate.