The short version: Since January 1, 2026, Ontario employers with 25 or more employees have been legally required to disclose AI use in every publicly advertised job posting, alongside a real pay range, honest vacancy status, and a promise to tell interviewed candidates the outcome within 45 days. This sits inside the Employment Standards Act, which means a non-compliant posting isn’t a style choice, it’s a complaint waiting to happen, with fines running up to $500,000 for repeat corporate violations. Nine months in, most of the employers actually using AI to screen or rank candidates still aren’t saying so in writing. If you post jobs publicly and use any tool that scores, ranks, or filters applicants, this is the one compliance item on this list that is cheapest to fix and most expensive to ignore.
What exactly does Ontario’s rule require?
Any employer with 25 or more employees that publicly advertises a job posting in Ontario must now disclose, in that posting, whether artificial intelligence is used to screen, assess, or select applicants. That sits alongside five other requirements introduced under the same reforms: posting a real expected pay range, confirming the position genuinely exists and hasn’t already been filled internally, dropping “Canadian experience required” language, telling interviewed candidates the outcome of their application within 45 days, and keeping copies of the posting and all related application materials for three years. (Workforce.com)
The definition of “AI” in the regulation is deliberately broad. It isn’t limited to a flashy generative-AI resume reviewer; a scoring algorithm, a keyword-ranking filter inside an applicant tracking system, or a tool that auto-rejects candidates below a threshold all plausibly qualify. If a piece of software is doing anything that changes which humans get looked at, the safer assumption is that it counts.
What happens if I don’t disclose it?
This lives inside the Employment Standards Act, 2000, which means it’s enforced the same way unpaid overtime or vacation pay violations are, through a Ministry of Labour complaint process, not a toothless registry. A rejected applicant who suspects AI screening was used and wasn’t disclosed can file a complaint directly. Penalties under Section 132 of the Act run up to $100,000 for a corporation’s first conviction, $250,000 for a second, and $500,000 for a third, on top of compliance orders. (SHRM)
Those ceiling numbers are for repeat, serious violations, not a guaranteed first-offense outcome, but the exposure is real in a way a lot of employers haven’t priced in. The compliance gap backs that up: industry reporting through 2026 has repeatedly found that most Ontario employers using AI in hiring aren’t disclosing it, which means the typical employer in scope right now is sitting on unrecognized risk, not a minor paperwork gap everyone has already closed. (DeployLabs)
Does this apply to my business if I’m small?
Only if you have 25 or more employees on the day you post a publicly advertised job. Below that threshold, the AI-disclosure and pay-transparency requirements in this specific regulation don’t apply to you yet. Two practical wrinkles matter here. First, “25 or more employees” is checked on the date of posting, so a business growing through that headcount needs to start complying with its very next public job ad, not retroactively fix old ones. Second, “publicly advertised” is the trigger, not internal promotions or postings shared only through private referral networks, so a 40-person company that only ever hires through word of mouth may not be in scope at all. If you’re close to the line and post jobs on Indeed, LinkedIn, or your own careers page, treat yourself as in scope rather than finding out the hard way.
What tools actually count as “AI” here?
Assume broadly, not narrowly. If your applicant tracking system auto-ranks resumes, auto-rejects based on keyword matches, scores video interviews, or recommends which applicants to interview first, that is very likely in scope even if your HR team never thought of it as “an AI tool,” it’s just the feature their ATS vendor turned on by default. Many popular applicant tracking platforms added ranking and scoring features over the last two years without employers necessarily realizing the regulatory weight that now attaches to using them. The fix isn’t to rip the feature out, it’s to say, in the posting, that it’s there.
What should I actually put in the posting?
The regulation requires disclosure, not a specific script, but a defensible version is a single sentence: “This employer uses artificial intelligence to assist in screening or assessing applications for this position.” That’s enough to meet the letter of the requirement without turning your job ad into a legal document. Pair it with the pay range requirement in the same posting, since both became mandatory at the same time and a posting that handles one but not the other is only half compliant.
What’s the actual cost of fixing this?
Close to zero for the disclosure line itself, it’s one sentence added to a template you already use for every posting. The real cost, if there is one, is the recordkeeping requirement: three years of posting and application records, and a 45-day clock on telling interviewed candidates the outcome. For a business running a handful of postings a year, that’s a folder and a calendar reminder, not a new system. For a business running dozens of postings a month, it’s worth confirming your ATS vendor already retains this data for three years by default, because several do not.
Frequently asked questions
Do I need a lawyer to write the disclosure language?
Not for the basic disclosure sentence itself; the regulation doesn’t prescribe specific wording. If your hiring process is complex or you’re unsure whether a specific tool counts as AI under the definition, a short consult is worth it, but most small and mid-sized employers can handle the posting-language fix internally.
What if I’m below 25 employees now but growing fast?
Start building the habit before you’re legally required to, since retrofitting compliance across every job template at the exact moment you cross the threshold is a worse time to do it than now, while there’s no deadline pressure.
Does this apply to postings on job boards like Indeed or LinkedIn, or just my own website?
It applies to any publicly advertised posting, which includes third-party job boards, not just your own careers page. If you post the same role in multiple places, each public posting needs to meet the requirement.
Can I get in trouble for a posting I made before January 1, 2026?
The requirements apply going forward from the effective date. A posting made and closed before January 1, 2026 isn’t retroactively non-compliant, but any posting still live or newly made after that date needs to meet the current rules.
What if my applicant tracking system doesn’t let me add this language easily?
Most ATS platforms allow custom text blocks in job templates. If yours genuinely doesn’t, the disclosure can go in the job description body itself rather than a dedicated field, as long as it’s visible to applicants viewing the public posting.
Get the next playbook in your inbox
Practical, no-hype AI guidance for small business owners. One useful email at a time, and you can unsubscribe anytime.
Does your job posting template currently disclose AI screening, or did this catch you off guard? Tell us in the comments.
