Ontario’s New AI Hiring Disclosure Law Has No Definitions Yet. That’s the Actual Risk.

Since January 1, 2026, employers in Ontario with 25 or more employees have had to disclose, on every publicly advertised job posting, whether artificial intelligence is used to screen, assess, or select applicants. That requirement now sits inside the Employment Standards Act, 2000 — not a new standalone AI statute, but an amendment to the same law that already governs overtime pay and termination notice. If your organization runs a GCC, delivery center, or BPO operation with an Ontario office, or if you sell screening technology to a Canadian employer, this is already in force. There is no phase-in period to plan around.

What the law actually says — and doesn’t

The disclosure obligation itself is narrow on paper: state, in the posting and any associated application form, whether AI is used to screen, assess, or select candidates. What makes this harder than it sounds is what the Act doesn’t define. “Artificial intelligence,” “screen,” “assess,” and “select” have no statutory definitions in the ESA amendment, and the province’s own regulations lean on an OECD-style definition — a machine-based system that infers from inputs to generate outputs — without settling how that applies to a resume-ranking plugin bolted onto an applicant tracking system, or a scheduling tool that happens to use a matching algorithm. The Ministry of Labour hasn’t issued detailed guidance despite the requirement already applying. Employers are being asked to comply with a law whose key terms nobody has authoritatively defined yet.

That ambiguity cuts in a specific direction for hiring teams that use any AI-assisted step in their funnel, not just an obvious “AI interview” product. A recruiter using an ATS with an AI-powered resume-ranking feature, a chatbot that pre-qualifies applicants, or a voice screening tool that auto-classifies candidates all plausibly trigger the disclosure requirement — and “plausibly” is doing a lot of work when there’s no enforcement guidance to check your interpretation against. Third-party tools compound the problem: many employers don’t know which parts of their existing recruiting stack already use AI under the ESA’s broad definition, because the AI is embedded in a platform they didn’t build and may not have audited.

Why this is a live question for Indian teams, not a Canadian one

We’ve written before about the four different legal theories US states are using to regulate AI hiring since the EEOC pulled its federal guidance — California’s vendor liability, Illinois’s private right of action, Texas’s intent standard, Colorado’s reasonable-care defense. Ontario’s approach is a different shape entirely: no liability theory to parse, no disparate-impact test, just a flat disclosure mandate baked into an employment standards law that already applies to every employer in the province. It’s closer in spirit to NYC’s Local Law 144 — transparency as the mechanism, not a fairness audit — except Ontario didn’t wait for an audit regime to mature before requiring the disclosure.

Several of the largest Indian IT-services and GCC operators run substantial Ontario-based teams — Toronto and the surrounding GTA is one of the more established delivery hubs outside India for firms with North American client bases. If your organization posts Ontario-based roles publicly, or if you’re a screening vendor whose Indian delivery team supports a Canadian employer’s hiring, the question isn’t whether this law applies eventually. It’s whether your current job postings and application forms already carry the disclosure line, and whether whoever wrote that line actually checked which parts of your funnel use AI as the ESA now defines it.

What to actually do before the Ministry issues guidance

Waiting for regulatory clarity is not a compliance strategy here, because the obligation is already live. The more defensible position is to audit your Ontario-facing hiring stack now: list every tool that touches screening, assessment, or selection, note which ones use AI in any form (including third-party features embedded in your ATS), and write a disclosure statement specific enough to survive scrutiny if the Ministry does eventually clarify the terms narrowly. A vague catch-all line (“this posting may use automated tools”) is a weaker position than a specific, accurate one, because it signals you didn’t actually check.

The uncomfortable part of this law isn’t the disclosure itself — most candidates already assume AI touches some part of a modern hiring funnel. It’s that Ontario chose to require honesty about it before defining what “it” means, which puts the burden of a defensible interpretation on the employer, not the regulator. That’s a harder position to be caught in than a law with clear rules and a grace period, because there’s no rulebook to point to if your interpretation turns out to be wrong.

If you’re screening candidates into Canada-facing roles from India and want to know what an audit-ready disclosure and screening record actually looks like, talk to us.

Sources: AI in Hiring: Ontario Employers Grappling with New Job Posting Disclosure Requirement — Osler, Hoskin & Harcourt LLP, New Year, New Rules: Ontario Job Posting Requirements Take Effect January 1, 2026 — Hicks Morley, A new Ontario law means your job search might look a little different in 2026 — CBC News