Job hunting in Canada is becoming a negotiation with software before a recruiter ever appears. In Ontario, new disclosure rules have made that shift unusually visible: covered employers must say when artificial intelligence is used to screen, assess or select applicants. Indeed Hiring Lab found AI-related language in 28% of Ontario job postings in May 2026, up from 9% in October 2025. The change is landing in a labour market where 1.5 million Canadians were unemployed in August and 24% had been searching for at least 27 weeks. For many applicants, an early round can now involve automated résumé screening, chatbots, prerecorded interviews or AI-assisted scoring. The convenience is real, but so are questions about fairness, privacy, accessibility and whether a human can explain why someone was screened out.
The First Interview May Happen Before a Recruiter Appears
A first-round interview no longer has to mean a scheduled call with a recruiter. One common format is the asynchronous, or on-demand, video interview: candidates receive prompts, record answers on their own time and submit them for later review. PwC Canada, for example, says its experienced-hire process uses on-demand video interviews before live interviews for many roles, illustrating how technology can sit between the application and the first human conversation.
That format should not automatically be called an “AI interview.” Recorded responses may be reviewed by people, scored by software, or both. Canada’s Public Service Commission draws the same distinction in its guidance, describing asynchronous interview platforms as an assessment method while separately discussing automated ranking and automated scoring. For job seekers, that distinction matters. A camera on screen may simply be recording an answer, or it may be part of a system generating scores and recommendations behind the scenes.
Ontario Has Pulled AI Hiring Into the Open
Ontario has pushed this issue into public view. Since January 1, 2026, employers covered by the province’s Employment Standards Act job-posting rules must state when they use artificial intelligence to screen, assess or select applicants for a publicly advertised position. The rules generally apply to employers with at least 25 employees on the day the posting appears, with defined exemptions for certain postings.
The disclosure is useful, but still limited. Ontario’s guidance says employers do not have to provide a detailed description of the AI system or explain exactly how it will be used; a statement that AI is involved can be enough. The same framework requires interviewed applicants to be told within 45 days whether a hiring decision has been made. That gives job applicants more visibility than before, but it does not necessarily reveal what data the system considered or how heavily an automated score influenced the outcome.
What the Software Can Actually Evaluate
AI can enter hiring long before a virtual interviewer begins asking questions. Federal staffing guidance describes systems that rank or sort résumés, generate assessment material, administer remote tests and automatically score tests or interviews. It also gives a cautionary example of video-interview software analyzing verbal answers, non-verbal communication and appearance before assigning scores or recommendations. Not every employer uses all of these features, but the technical range is very broad.
Canada has also seen public-sector experimentation. A Treasury Board privacy-impact summary described a pilot using Knockri, an AI hiring platform, for portions of an Administrative Services and Executive staffing process. The platform was described as using natural-language processing and machine learning to analyze interview responses against job-related criteria. For an applicant, the practical lesson is simple: “AI in hiring” can mean anything from sorting an application to evaluating recorded answers, so the disclosure alone may not tell the full story.
Why Employers Are Automating the First Round
Employers have operational reasons to automate the first stage. Canada’s federal video-recruitment toolkit lists fewer scheduling problems, fewer time-zone conflicts, reduced travel, less administrative work and the ability to handle applicant pools nationally as advantages of prerecorded interviews. An employer can ask questions of every candidate and let multiple reviewers assess responses without finding one time when everyone is available.
Research on asynchronous video interviews points to that trade-off. A Human Resource Management Review paper notes that these interviews can be faster, cheaper and less demanding of employee time, while platforms are designed to reduce scheduling burdens and expand the number of applicants screened. For a national employer hiring from Halifax to Vancouver, that efficiency is very easy to understand. What saves hours for a recruiting team, however, can shift responsibility onto candidates, who must manage the technology, environment and communication format without the cues of a live conversation.
The Candidate Experience Can Feel Colder
Efficiency does not guarantee a good experience. In a 2019 experiment with 180 observations, applicants viewed synchronous video interviews more favourably than asynchronous ones, although fairness perceptions did not significantly differ between human and AI decision agents. Another experiment involving 148 participants found highly automated interviews were seen as more consistent but as providing less social presence, and that lower social presence reduced organizational attractiveness.
That finding matches the awkwardness of speaking into a camera without receiving a nod, follow-up question or sign that an answer landed well. The absence of a live interviewer removes interpersonal signals that help a candidate pace an answer or clarify a misunderstanding. Other research has found algorithm-based evaluation can reduce perceived fairness and feedback acceptance compared with recruiter-based evaluation. Employers may gain consistency, but poorly explained automation can make hiring feel distant at the moment a candidate is trying to make a human impression.
Bias Does Not Disappear When the Interviewer Is Software
Replacing a recruiter with software does not make discrimination disappear. The Ontario Human Rights Commission has warned that AI systems can reproduce patterns found in historical hiring data, use proxies such as postal codes and employment gaps, and create barriers for groups protected by human-rights law. Its submission on Ontario’s AI disclosure rule cited a Peel District School Board review in which an algorithmic screening tool had inappropriately filtered out qualified racialized teaching candidates.
The principle is straightforward: employers are still responsible for discriminatory outcomes when a vendor or algorithm helps produce them. Ontario’s Human Rights Code protects employment as a social area, while federally regulated employment is also subject to the Canadian Human Rights Act. That is why transparency is only a safeguard. Testing, job-related criteria, monitoring and a way to challenge questionable results matter. Automation can standardize decisions, but standardized bias is still bias—and it can scale.
Accessibility Is a Test of the System, Not the Applicant
Accessibility may be a clear test of whether an automated interview is measuring the job or comfort with the technology. The Ontario Human Rights Commission has warned that interview technologies can be unreliable for people with speech impediments, candidates who use screen readers and people whose first language differs from its training data. A rigid chatbot can also mishandle an accommodation request that a recruiter might recognize.
Accessibility Standards Canada’s national employment standard sets an important national benchmark for organizations using applicant-tracking systems and AI. It says screening should focus on bona fide occupational requirements, organizations should show their systems are not discriminatory, and candidates should receive information about accommodations and how AI is used. The standard is not a blanket hiring law for every Canadian employer, but it captures an important principle: a system should test the ability to do the work, not unrelated sensory, speaking or technical characteristics.
Recorded Interviews Create a Bigger Privacy Footprint
Recorded interviews create more privacy risk than a conventional conversation. The Office of the Privacy Commissioner of Canada has warned federal institutions that asynchronous staffing platforms can collect a candidate’s answers plus their likeness, voice biometric and anything visible in the camera frame. A recording can be replayed, retained and accessed repeatedly, often through a third-party recruitment platform.
That changes what Canadian applicants may reasonably want to ask today. Who owns the recording? How long will it be kept? Who can view it? Is the video analyzed, or only the transcript? Federal privacy guidance recommends limiting collection, reviewing third-party retention practices and helping candidates protect unnecessary background information. It has advised candidates to use a neutral space or background blur when appropriate. A bookshelf, family photo or medical device may have nothing to do with job performance, yet a recording can easily capture it permanently unless the process is designed carefully today.
Preparation Still Matters, but Gaming the Algorithm Is Risky
Candidates can prepare for automated interviews without reverse-engineering an invisible algorithm. A 2023 Journal of Vocational Behavior study from Saint Mary’s University and the University of Calgary found that short training improved asynchronous-video interview performance and perceptions of consistency. In one study, 202 participants completed mock interviews; a second included 156 active job seekers. Training was associated with more structured responses, while practice alone had limited effects.
A more practical preparation approach follows: prepare concise examples, organize answers around a clear situation, action and result, test the camera and microphone, and read all instructions carefully. PwC Canada tells candidates using its on-demand process to check lighting, background, internet connection and equipment before recording. Candidates should look for AI disclosures and accommodation instructions rather than guessing what software might measure. When the process is opaque, asking whether answers are automatically scored can be more useful than performing for an imagined algorithm.
Human Accountability Is Becoming the Bigger Question
Canada’s strongest governance model still keeps a person accountable for the result. In federal public-service hiring, managers using AI must understand their systems, validate outputs and able to explain decisions. When AI recommends or supports a staffing decision, federal guidance requires an Algorithmic Impact Assessment, candidate notice, a meaningful explanation of how and why the decision was made, and a process for raising concerns or challenging the outcome.
That standard goes further than Ontario’s private-sector posting rule, which focuses on disclosure rather than a detailed explanation of the algorithm. The direction is clearer: hiring steps can be automated, but transparency pressure and human oversight are rising with them. For job seekers, the first “interviewer” may increasingly be a platform, chatbot or scoring model. The question is not whether software participates in hiring, but whether an employer can show the system is relevant, fair, accessible and answerable to a human decision-maker.
