AI is creeping into Canadian workplace disputes. Is it just creating more work for HR?

Artificial intelligence was sold to employers as a way to tame paperwork, spot risks early and ease the load on HR teams.

But in Canada, AI is starting to appear in legal dust-ups – from self-represented parties submitting AI-generated “research” to regulators warning that algorithmic hiring tools could land employers in front of a human rights tribunal.

For HR leaders, the question is becoming hard to avoid: is AI improving fairness in the workplace, or simply generating a new category of avoidable claims and administrative work?

A Quebec cautionary tale: AI hallucinations in court

The starkest warning so far comes from Quebec.

In a recent case before the Quebec Superior Court, a self-represented litigant was fined C$5,000 after he filed submissions packed with non-existent cases and bogus citations generated by an AI tool. The judge described the conduct as “highly reprehensible” and a serious breach of procedure, stressing that AI output must be rigorously checked by humans before it ever reaches a courtroom.

Read next: Lack of leader communication fuelling AI anxieties, report finds

The dispute itself had nothing to do with employment. But the behaviour – a non-lawyer asking an AI to build their case, then dropping the result into formal documents – is exactly what employment and human-rights lawyers say they are now seeing in workplace disputes.

Canadian employment firms are reporting clients who use ChatGPT to draft settlement responses, employment contracts and severance calculations, often based on templates or formulas that don’t match Canadian law or the correct province. Others warn that AI tools offer none of the accountability, nuance or duty of care that come with real legal advice – and that when things go wrong, the consequences fall squarely on the user.

For HR, that means more time spent untangling AI-drafted demand letters, correcting misconceptions about notice and severance, and responding to lengthy complaints that sound authoritative but rest on shaky foundations.

AI in hiring: discrimination risk by design

If generative AI is complicating the claims side, algorithmic tools are raising questions at the front end of the employment relationship.

Across Canada, more employers are experimenting with AI-enabled recruitment platforms, résumé screeners and video-interview analytics. These systems promise faster shortlisting and better “fit” – but they also risk baking in bias from historical data or flawed design.

Recognising this, Canadian human-rights bodies and law-reform organisations have developed an AI impact-assessment framework specifically aimed at identifying discrimination risks. It is designed to be used before and after deploying AI systems in areas such as hiring, promotion and performance management, and it squarely targets both public and private-sector employers.

Read next: Why HR leaders can't be afraid to talk about AI

At the same time, Canadian HR and legal commentators are warning that AI in hiring can easily fall foul of provincial human-rights codes if it disadvantages candidates on the basis of race, disability, age, gender or other protected grounds. Systemic discrimination is a real risk where algorithms simply replicate the biases embedded in their training data – and in practice it is employers, not vendors, who are likely to be held responsible.

Layer onto this growing concern about accent bias. Canadian commentary has highlighted how “foreign-sounding” accents can unconsciously affect hiring decisions, especially when recruiters or tools place excessive weight on fluency and “clarity” rather than skills. If AI-powered video-interview tools struggle to accurately process accented speech or atypical communication styles – a problem already documented overseas – HR may find itself defending not just an individual decision, but the entire system that produced it.

More forums, more complexity

Canada’s legal landscape gives employees several paths to raise concerns:

• human-rights complaints at federal or provincial commissions

• claims under employment-standards legislation

• union grievances

• civil actions for wrongful dismissal.

A single decision influenced by AI – say, a rejected candidate, a failed promotion, or a selection for redundancy – can therefore spawn multiple overlapping challenges.

Read next: GenAI spending to go up as adoption enters new phase

Legal commentators already note that the “multi-forum” nature of Canadian employment law is driving up the volume and complexity of disputes even without AI in the mix. Add opaque algorithms to that environment, and HR may be facing requests for technical documentation, vendor contracts and data-audit material on top of the usual interview notes and performance records.

Is AI actually saving HR any time?

On paper, AI should help HR teams:

• triage résumés

• identify potential risks in large workforces

• generate first-draft policies and letters

• surface patterns in grievances or turnover.

Some of that is happening. But the early legal signals suggest an uncomfortable trade-off:

Every hour saved by an automated screener or AI-drafted memo may be matched by extra hours spent dealing with AI-fuelled misconceptions, governance requirements or complaints.

Three dynamics in particular are starting to create extra work:

1. AI-written claims and demand letters
   Employees and former employees can now produce polished-looking documents in minutes. HR then has to review them carefully, involve legal, and respond, even when the underlying case is weak or out of time. That is not time HR would have spent if the individual had been forced to write the claim unaided.

2. Discovery fights about algorithms
   Where AI has influenced a decision, unions and lawyers will naturally ask: How does the system work? What data was used? Was it audited for bias? Gathering those answers – sometimes from global vendors – is squarely an HR and legal task.

3. Policy and training obligations
   Impact-assessment tools and human-rights guidance are valuable, but they also imply ongoing work: documenting uses of AI, running assessments, recording mitigation steps and training managers on what the tools can and cannot do.

All of this prompts a fair question for Canadian HR leaders: are you actually reducing your workload with AI, or simply shifting it into new – and riskier – forms?

What HR can do now

The answer is not to swear off AI entirely. Used carefully, it can still help with volume tasks and pattern-spotting. But the Canadian experience so far points to a few practical guardrails:

• Treat employment-related AI as “high-risk” technology. Assume that hiring and HR tools will be scrutinised through a human-rights lens. Involve legal and privacy teams early, and use structured impact-assessment frameworks rather than relying on vendor marketing.

• Insist on explainability. If a system cannot provide intelligible reasons for its rankings or scores, think twice about using it in recruitment, promotion or termination decisions.

• Keep humans clearly in charge. Make sure every adverse decision – not shortlisted, not promoted, selected for layoff – can be explained on human reasoning alone, with AI as a secondary input, not the decider.

• Prepare for AI-drafted claims. Train HR teams to recognise when letters or complaints are likely AI-generated and to sanity-check the legal assertions before reacting. When in doubt, get early advice rather than engaging in protracted back-and-forth on flawed assumptions.

• Educate employees and managers. Make it clear that AI is a tool, not a lawyer – for both sides. Encouraging staff to seek real legal or HR advice early may head off the kind of misadventures seen in Quebec.

The open question

Unlike Australia, where  the Fair Work  Commission is already ringing alarm bells  over a rising tide of AI induced claims, Canada has not yet seen a wave of tribunal decisions explicitly blaming AI for breaches of employment or human-rights law. There are no official statistics tying AI use to a spike in complaints. For now, the evidence is anecdotal but consistent: judges sanctioning AI misuse, human-rights bodies issuing new guidance, and law firms warning that they are spending more time unwinding problems that started with a chatbot.

For HR managers, that leaves a live question hanging over every new AI pilot:

Is this tool genuinely helping us build a fairer, more efficient workplace – or are we quietly signing up for the next round of grievances, document requests and hearing days?

The technology is not going away. But Canadian employers still have a choice: deploy AI in HR cautiously, with human rights and governance front and centre – or let the courts and commissions, eventually, answer that question for them.