Penalties have doubled. The Ministry still hasn’t issued guidance on key definitions
Canada’s new national AI strategy, announced earlier today, promises to modernize federal privacy legislation and introduce online safety laws. Those reforms have no firm legislative timeline. The compliance obligation your organization faces today is not future-looking. It has been in force since January 1, 2026, it is enforceable now, and the financial consequences of getting it wrong have doubled.
READ MORE: Federal government releases national AI strategy putting workforce transformation at the centre
Under Ontario’s Working for Workers Four Act, 2024, every employer with 25 or more employees must disclose in any publicly advertised job posting whether artificial intelligence is used to screen, assess, or select applicants. The same disclosure must appear in any associated application form. Under the companion Working for Workers Five Act, penalties for Employment Standards Act offenses have increased from $50,000 to $100,000.
“The penalties for ESA offenses have doubled from $50,000 to $100,000,” Michelle Cook, an associate lawyer at Aird & Berlis in Toronto, told HRD Canada in our January compliance briefing, warning that organizations “can’t treat these changes as a minor policy update.”
Five months later, many still are.
The guidance gap making compliance harder than it should be
Here’s the problem no employer talking point adequately addresses: the Ontario Ministry of Labour has not yet issued definitive guidance on what the law actually requires in practice.
Osler, Hoskin & Harcourt noted in its March 2026 employment law analysis that the Ministry “has yet to issue guidance on what is captured by the broad definition of ‘artificial intelligence,’ how to interpret key terms such as ‘screen,’ ‘assess,’ or ‘select’, or the level of detail required when disclosing the use of AI.” That ambiguity puts employers in the uncomfortable position of needing to comply with a law whose precise scope remains unsettled.
READ MORE: Can governments ‘heavily regulate' AI and tech companies?
The definition of AI in the legislation is deliberately broad. It captures dedicated applicant tracking system scoring modules, resume screening tools, automated ranking engines, and chatbot-based initial candidate screening. If your organization uses any commercially available recruiting platform such as Workday, Greenhouse, Lever, or iCIMS, the question isn’t whether that tool uses AI. It almost certainly does. The question is whether you’ve disclosed that use in every public job posting and application form since January 1.
Employment lawyer Paulette Haynes, founder of Haynes Law Firm in Toronto, has flagged a human rights dimension that many compliance reviews miss. As our earlier analysis of AI hiring risks under Ontario’s Human Rights Code noted, AI hiring tools can inadvertently perpetuate or amplify biases embedded in their training data, raising the risk of systemic discrimination. The disclosure obligation is the floor, not the ceiling.
The inclusion of AI disclosure “further complicates hiring workflows,” employment lawyer Ilinca Fusco told HRD Canada. “HR and legal counsel will need to work closely to develop these disclosure statements for candidates in the job postings and potentially develop an AI policy.”
Five things your organization needs to have in place
1. Audit your full technology stack before your next posting goes live.
List every tool in your recruiting workflow, including ATS, resume screening, video interview scoring, skills assessments, and chatbot pre-screening, and determine whether any use AI at any stage of candidate screening, assessment, or selection. The obligation is triggered by AI use at any point, not only at the final decision stage. If you’re uncertain whether a tool qualifies, assume that it does. Consult the vendor’s documentation and, if necessary, legal counsel.
2. Update every public job posting and application form and lock the disclosure into your template.
Your disclosure doesn’t need to be technically detailed, but it must be accurate and present in every relevant document. Practical language used by compliant organizations includes variants such as: “We use artificial intelligence tools to assist in screening and assessing applications for this role.” Lock the disclosure into your ATS posting template so postings can’t be published without it. Tailor the language to what your tools actually do and avoid generic language that overstates or understates actual AI involvement.
3. Implement the 45-day candidate notification rule at the same time.
The legislation introduced a second obligation that many organizations have been slower to address: employers with 25 or more employees must notify every interviewed candidate whether a hiring decision has been made, within 45 days of their most recent interview. As HRD Canada reported when this rule was introduced, failure to comply can lead to Ministry of Labour complaints and enforcement action. Records of these notifications must be retained for three years, alongside copies of every publicly advertised job posting and associated application form.
4. Train your HR and recruiting teams, including how to answer candidate questions.
Ontario’s own audit of its public sector workforce, published earlier in 2026, found that of approximately 55,000 Ontario Public Service staff, just three percent had completed the Ministry’s Responsible Use of AI training course, and that wasn’t even a mandatory program. Private sector organizations that haven’t trained their HR and recruiting teams on what the disclosure obligation requires, and how to respond if a candidate asks specifically which AI tools are being used, are carrying reputational and legal exposure that a focused training session could substantially reduce.
5. Understand that federal changes will layer on top of, not replace, what Ontario already requires.
The Carney strategy’s promise to modernize PIPEDA and introduce online safety laws will, when enacted, add obligations on top of provincial requirements. Quebec’s Law 25 is already fully in force and imposes formal privacy impact assessment obligations for AI systems that significantly affect individuals. British Columbia will require pay range disclosure in postings from November 2026. The compliance landscape is layering, not consolidating, and as our comprehensive guide to Ontario’s new pay transparency and AI disclosure laws made clear when the rules were first announced, organizations that treat this as a one-time posting update rather than an ongoing governance obligation will find themselves repeatedly exposed.
What the Carney strategy signals for HR compliance planning
The strategy’s commitment to create a Canada Trusted AI Certification Program, designed to help Canadians identify trustworthy AI products, isn’t yet a mandatory standard. But its direction is clear: AI tools used in consequential decisions, including hiring, will face increasing transparency and certification pressure at the federal level. For HR leaders currently procuring AI-powered recruiting platforms, the practical question is whether the tools purchased today will be compatible with a federal certification regime that may arrive within the current contract term.
The most urgent action for HR compliance teams isn’t to wait for that guidance. It’s to ensure that every job posting published today meets the standards Ontario law already requires, with the doubled penalties that now attach to falling short.
