Ontario's Working for Workers Act, 2025 came into force on January 1, 2026. One of its requirements — that employers disclose the use of artificial intelligence in the hiring process — has been sitting in job posting templates across the province since then, mostly unadded.

This is not a compliance technicality buried in a schedule somewhere. It is a candidate-facing obligation with ESA enforcement behind it. If your organization has posted any jobs in Ontario since January 1 without the disclosure, those postings are non-compliant. That is not a hypothetical exposure. Ministry of Labour inspections under Working for Workers provisions have been active since the Act's earlier amendments, and this requirement is specific enough to check.

Most HR teams I speak to fall into one of two categories. The first has added a vague line about "technology" to their postings and called it done. The second has not addressed it at all because it landed in the gap between HR, recruiting, legal, and the vendor relationship — and no one owned it clearly enough to act.

Neither category is where you want to be.

What the law actually says

The requirement under the Working for Workers Act, 2025 is straightforward in its language: employers that use artificial intelligence in the hiring process must disclose this to candidates in the job posting. The disclosure obligation is on the employer, not the vendor. If your ATS vendor is doing the screening, you are the one required to tell candidates that AI is being used.

The definition of AI under the Act is functional rather than technical. The question is not whether your vendor uses the word "AI" in their marketing. The question is whether the tool makes automated decisions about which candidates advance, scores or ranks applications before a human reviews them, or filters submissions based on algorithmic criteria. If the answer is yes to any of these, the disclosure requirement applies.

This catches a wider range of platforms than most HR teams realize. Greenhouse, Lever, Workday Recruiting, iCIMS, Ceridian Dayforce, BambooHR — most of these platforms have AI-assisted screening features that are enabled by default. Many organizations set them up years ago and have not revisited the configuration since. The features are running in the background, shaping which candidates a recruiter ever sees, and the candidates have no idea.

That is precisely the gap the law is designed to close.

Why vague language does not work

A number of employers have responded to this requirement by adding language along the lines of: "We may use technology to support our hiring process." This does not satisfy the obligation.

The Ministry has not published a prescribed form for the disclosure, but the intent of the Act is transparency about AI specifically, not technology in general. "Technology" covers email. It covers spreadsheets. It covers the phone call your recruiter makes from their smartphone. The disclosure requirement exists because AI-driven screening is categorically different from other hiring tools — it makes automated judgments about candidates without human intervention at the point of decision. That is what candidates have a right to know about.

A compliant disclosure names AI or artificial intelligence specifically and connects it to the hiring process. One sentence is sufficient:

"We use artificial intelligence tools to assist with the screening and evaluation of applications as part of our hiring process."

That sentence added to every job posting template closes the compliance gap at the posting level. It takes about four minutes to implement. The reason most organizations have not done it is not that it is difficult. It is that it requires someone to actually decide it is their job to do it.

The vendor knowledge gap

Here is the part that creates the most risk for HR teams who think they are already compliant.

Most people who selected and configured their organization's ATS did so several years ago. The platforms have changed significantly since then. Vendors have added AI features, enabled them by default, and released them through product updates without requiring administrators to opt in. If you have not reviewed your ATS configuration recently, there is a reasonable chance your platform is doing things algorithmically that you are not aware of and have not disclosed.

The specific features to ask about are: automated resume screening or parsing that scores or ranks candidates, fit scores or match percentages generated before human review, keyword-based filters that eliminate candidates before a recruiter sees the application, and any predictive analytics applied to the candidate pool.

Request this documentation from your vendor in writing. Ask specifically which features in your current configuration use AI or machine learning. Ask whether any of these features are enabled by default. Document the response. If enforcement comes, that documentation demonstrates good-faith inquiry. If the vendor cannot answer clearly, that is a gap worth escalating internally.

What the enforcement landscape looks like

The Ministry of Labour enforces the ESA through workplace inspections, complaints from employees, and proactive review campaigns. Working for Workers provisions have been a priority enforcement area since the Act's earlier amendments on disconnecting from work and non-compete clauses.

The AI disclosure requirement is specific, checkable, and visible in the job posting itself. An inspector does not need to come into your office to identify non-compliance. They can look at your job postings.

The consequences for non-compliance include orders to comply, and escalating consequences for continued failure to comply. For organizations that have been posting jobs since January without the disclosure, the practical first step is remediation with documentation of when the correction was made, not hoping the issue resolves itself.

The broader picture

Ontario's AI disclosure requirement is the first of what will be several Canadian legislative responses to AI in hiring. Quebec has been watching closely. British Columbia has had preliminary consultations on similar provisions. Federal regulated employers are widely expected to see requirements under the Canada Labour Code in the next legislative cycle.

The employers building clean, documented AI disclosure practices now will not be scrambling when the next amendment lands. More importantly, candidates are paying attention in ways they were not two years ago. The discourse around algorithmic hiring, automated rejection, and AI bias in recruiting has moved from niche tech media to mainstream coverage. An employer that proactively discloses AI use and explains how it is used in the hiring process is differentiating itself in a candidate market that is increasingly skeptical of opaque screening processes.

This is compliance that also happens to be good practice.

What to do this week

Update your job posting template with a clear, specific AI disclosure. Document the date you made the change. Review your ATS configuration or request vendor documentation to confirm which features are active. Assign a single owner for this across HR, recruiting, and legal so the disclosure stays current as your tools evolve.

If you use multiple job posting platforms — LinkedIn, Indeed, Workopolis, your own careers page — confirm the disclosure appears on all of them, not just in your internal template.

The window for quiet remediation is open now. It is easier to address this proactively than to explain it after an inspection.

Not legal advice. Questions about your specific ATS configuration and disclosure obligations should be reviewed with employment counsel familiar with Ontario's Working for Workers Act.