AI, people risk and employment law: What Canadian employers need to know in 2026

AI is changing how fast work gets done. It isn’t changing who’s accountable when a people decision goes sideways. That was the throughline of our recent webinar with Jay Spiro, Senior Legal Consultant at Vistera, and Nevena Radoicic, Partner Manager, who unpacked the employment law trends every Canadian business leader should have on their radar heading into the back half of 2026.

The reassuring part? Most legal risk still comes down to the fundamentals: leadership, communication and documentation. Those haven’t changed much. The trick is applying them well in a workplace that’s shifting faster than ever. Jay framed the whole conversation around three questions every employer should be asking: Are we making good people decisions? Are we managing changing expectations properly? And are we using AI responsibly?

Here’s what we learned.

Want the full session, case law breakdowns and live Q&A included?

Lens one: Making good people decisions

Jay opened with a truth that catches a lot of businesses off guard. Most employment law problems don’t start as legal problems. They start as management problems that quietly escalate.

The pattern is almost always the same. A concern comes up. There’s an informal chat. Follow-up never quite happens, documentation stays thin, and then one day the employee is blindsided by a management action they didn’t see coming. A dispute follows. Termination is where the problem becomes visible, but it’s rarely where it began.

Documentation is the theme that never goes away

If there’s one issue that shows up in employment disputes over and over, it’s documentation. Not because managers aren’t having conversations. They usually are, in person, over Teams, through Slack, by text. The problem is that six months later, nobody can prove what was actually said, when it was said or what expectations were set.

Jay offered a brilliantly simple test:

Would the file make sense to someone reading it six months from now?

If a manager, an HR lead, a lawyer or a tribunal member picked up that personnel file, could they follow the story? Strong documentation tells that story. And this is exactly where AI earns its keep. Teams recaps, meeting summaries and transcript tools can turn a conversation into clear, shareable notes in seconds.

But Jay drew a sharp line: AI can summarize a conversation. It can’t replace one. It won’t conjure a discussion that never happened. And once you’ve documented something, it needs to live somewhere secure, with proper privacy and access controls. This is where a centralized employment platform beats a manager’s notebook every time. Performance discussions, manager notes, goals and follow-ups all sit in one place, with permissions you actually control. The bottom line? Good documentation doesn’t just help you defend a claim. More often, it stops the claim from ever arising.

The performance management and mental health overlap

One of the fastest-growing risk areas Jay is seeing is the collision between performance management and mental health leave. The scenario goes like this: an employee is underperforming, a manager raises it and within a few conversations the employee becomes upset and starts a stress leave.

Suddenly, employers are asking nervous questions. Can we keep managing performance? Should we go completely hands-off? Are we exposed to a human rights claim if we push? Jay’s core point: a legitimate performance issue doesn’t vanish because a leave begins. What matters is proving the issue existed before the leave, that expectations were communicated and that coaching and support were offered. When documentation only starts after the leave, employers get hit with allegations that the concerns were invented in response to the leave or the disability.

Most of the time, that accusation isn’t true. The employer simply didn’t document clearly enough beforehand, which brings it all back to consistent record-keeping, showing escalation and training your managers to do this well. Managers are your first and best line of risk prevention. They can also be your biggest risk.

Lens two: managing changing workplace expectations

Flexibility has rewired what employees expect from work. Hybrid setups, remote arrangements, flexible schedules. People want autonomy over how their day runs. At the same time, many employees still crave clarity. They want to know where their role starts and stops and what success looks like. Employers, meanwhile, are often chasing productivity, culture and collaboration, and they increasingly need people who can adapt as AI reshapes their duties. So you get a tension: employees wanting flexible arrangements but certain duties, employers offering flexible arrangements but expecting flexible duties. Neither side is wrong. But friction breeds legal risk. This is where thoughtful workforce planning starts to matter, because how you structure roles today shapes your exposure tomorrow.

Two constructive dismissal cases worth knowing

Jay walked through two recent decisions that landed in the same place.

• British Columbia Court of Appeal: An 18-year employee had worked remotely for a long stretch. The employer directed a full-time return to the office. The employee refused and claimed constructive dismissal. The court agreed the remote arrangement had become a permanent term of employment, so demanding an immediate full-time return was a constructive dismissal.
• Ontario, 2025: An employer let an employee work remotely from Europe for over a year, then required a return and significantly cut their responsibilities. The employee claimed constructive dismissal. The court agreed again.

A quick refresher: constructive dismissal is when a court finds employment was effectively terminated by the employer, even though the employer never expressly said so. Here’s the part people get wrong. The lesson isn’t that once you allow remote work, it’s locked in forever. The real lesson is that long-standing arrangements matter. Let something run for months or years without defining expectations or documenting that it’s temporary, and it can quietly become a contractual term even with nothing in writing.

The risk also climbs when changes stack. A return-to-office mandate on its own might be manageable. A duties change on its own might be fine. Bundle several big changes together, and the constructive dismissal risk jumps sharply. Jay’s takeaway was refreshingly clear: flexibility is fine. Flexibility should be documented. The same logic applies to shifting someone from salary to hourly, or hourly to salary. If it runs long enough without documentation that it’s discretionary and temporary, it becomes a core term of employment. You can still change it, but only with the employee’s consent or with proper advance notice.

That word, notice, came up again and again. Employers often ruminate for months over bringing people back, then spring the decision on staff who are hearing it for the first time. Do it with consent, or do it with notice. Ambush is where lawsuits live.

Ready to bring your people processes into one place, so documentation and role clarity are built in from day one?

Lens three: using AI responsibly

AI is already woven through HR. Job ads, resume screening, candidate ranking, performance reviews, policy drafting, investigation summaries. And whether it’s officially approved or not, your employees are already using it. A live poll during the webinar showed only seven percent of attendees weren’t using AI at all, with most using it informally. Our own AI Paradox Report reveals just how widely and quietly this shift is unfolding across Canadian workplaces.

So the question is no longer whether AI is being used. It’s whether it’s being used responsibly.

Regulation is already here

As of January 1 this year, Ontario employers with 25 or more employees must disclose when AI is used to screen, assess or select candidates in publicly advertised job postings. Jay sees this as a signal of what’s coming: more rules built around transparency, accountability and explainability.

For now, and for the foreseeable future, accountability sits with the people using the tools, not the developers who built them. So a practical gut check: can your organization explain how AI influences your hiring decisions?

Four AI risks every employer should understand

Jay laid out four that deserve real attention:

• Confidential information in public AI tools. Accommodation requests, medical details, complaint records, investigation notes. Do your people know what can and can’t be entered into an AI tool, and which tools are secure?
• Unreviewed AI-generated performance reviews. The danger isn’t that AI gets something wrong. It’s that a busy manager assumes it’s right and never verifies it.
• Over-reliance on automated hiring. You might be trying to strip out human bias in ranking and screening, only to bake in the AI’s own hidden biases. You still need to know what it’s using to make those calls.
• AI-generated policies. An AI-drafted handbook can look polished and professional and still fail to reflect your actual workplace. Slick output isn’t the same as accurate output.

The message tying them together: using AI doesn’t transfer accountability to the AI. The employer stays responsible.

The real opportunity: AI plus human expertise

Jay pushed back on the framing that gets thrown around constantly. AI versus lawyers. AI versus HR. He argued the biggest opportunity isn’t replacing expertise at all. It’s improving access to it.

Most small businesses and non-profits aren’t short on information. They’re drowning in it. And workplace problems, a termination, a harassment complaint — never arrive at a convenient moment. When you pair AI with human experts, you can spot issues faster, get quicker answers and help managers make better-informed decisions, without losing the judgment only a person brings. AI accelerates. It doesn’t sit accountable.

The organizations that manage people risk best in 2026 won’t just be the ones that adopt AI. They’ll be the ones combining smart technology with good leadership, sound judgment and access to trusted expertise.

Your 2026 employer checklist

Jay closed with a practical checklist for the rest of the year.

Better decisions

• Review what your manager training actually looks like
• Improve documentation across the board
• Address issues early, before they escalate
• Strengthen your accommodation processes

Better expectations

• Review remote work arrangements and the flexibility that’s crept in
• Decide what’s permanent and what stays flexible
• Communicate any changes carefully and well in advance

Better AI governance

• Put a clear AI use policy in place
• Understand how AI is being used and protect confidential information
• Create guidelines so employees aren’t tiptoeing around the tools
• Train your team and approve a set of trusted tools

None of this requires ripping up how you work. It’s about applying skills you already have to a workplace that’s moving quickly, and giving your managers the tools and clarity to make good calls under pressure.

The businesses that get this right treat people risk as something to manage proactively, not firefight after the fact. The right systems make that far easier, keeping your documentation, records and processes connected in one place instead of scattered across notebooks, inboxes and spreadsheets.

Want to see how Employment Hero helps you document decisions, manage expectations and stay ahead of people risk?