IBM's Institute for Business Value surveyed 1,500 CHROs and equivalent executives, plus 8,800 employees, and released the Canadian findings on October 2, 2026. In Canada, 37% of organizations don't involve the CHRO when they define AI strategy. And 87% of Canadian CHROs believe AI adoption creates "invisible" work for employees, like validating recommendations, fixing mistakes, providing context and managing exceptions.
That 87% tells me there's a gap between adopting AI and knowing how to use it. We're all still figuring out how to use it effectively, and most organizations haven't caught up.
In hiring, I'd give AI as much of the heavy lifting as it can carry and keep people accountable for every outcome. I also think AI is going to create more jobs. I worry about organizations that bolt it onto the process they already run without rethinking how the work gets done or building the skills their people need to work alongside it.

How Does AI-Generated Invisible Work Impact Recruiter Productivity and Salary Costs?
I run my team on a formula: performance equals potential minus interference. So we've always asked what someone needs to do the job, whether that's two screens, an ergonomic chair, money for advertising or a credit card. AI should be making people more productive, so it belongs on that list as one more way to take friction out, right? The checking and fixing those CHROs describe is interference, and a tool that piles it onto a recruiter's day subtracts from performance.
If checking and fixing AI output eats an hour a day, that's about 250 hours a year. A recruiter on $100,000 costs about $50 an hour over 2,000 working hours, so those 250 hours come to roughly $12,500 in salary. I estimate every employee should return three to four times their salary, which puts the lost output at $37,500 to $50,000 a year per recruiter.

So I'd hold an AI tool to the same standard as the chair or the second screen. Is it taking work off your recruiters or handing them new work? Who checks what it produces, and how long does that take? If a sales tool cost your reps an hour a day, would you renew it?
In IBM's 2026 CHRO study, Canadian CHROs were more likely than CHROs globally to say AI creates invisible work, while fewer Canadian organizations leave the CHRO out of AI strategy.
|
IBM 2026 CHRO study finding |
Canada (75 executives, 452 employees) |
Global (1,500 executives, 8,800 employees) |
|---|---|---|
|
CHROs who say AI adoption creates invisible work for employees (validating, fixing mistakes, providing context, managing exceptions) |
87% |
|
|
Organizations that don't involve the CHRO in defining AI strategy |
37% |
46% |
|
Organizations that clearly define which work is human-led, AI-assisted or AI-executed |
Not reported |
26% |
|
Employees who feel safe questioning or overriding AI when the CHRO at least shares responsibility for which decisions stay human-led |
Not reported |
|
|
Employees who feel safe questioning or overriding AI when HR is only advisory |
Not reported |
43% |
|
CHROs who name supervising, validating and overriding AI outputs as the workforce's most essential skill |
68% |
71% |
Which Candidate Sourcing and Recruiting Tasks Should Be AI-Assisted Versus Human-Led?

Globally, IBM found 76% of employees feel safe questioning or overriding AI recommendations where the CHRO at least shares responsibility for which decisions stay human-led. Where HR is merely advisory, it's 43%. I read that as people wanting a human to make the final call. The "advisory" in IBM's finding is about HR's role, though. For AI in hiring, I think advisory is too limiting.
Only 26% of organizations globally clearly define which activities are human-led, AI-assisted or AI-executed, according to IBM. In recruiting, the first thing I'd hand to AI is sourcing and research. It should understand the role, research the market, identify potential candidates, assess their experience against the requirements and help prioritize who to speak to, so hiring gets faster and decisions get made on better data.
The recruiter owns the quality, the conversation, the relationship and the gut read, and makes the final call, because judgement and understanding people matter so much in hiring. Top performers are less likely to answer a bot than a person they know, especially when they're weighing a big career move. It's like hockey scouting, where the stats narrow down the free agents and the GM owns the signing.
That recruiter has to be free to overrule the AI's ranking. A ranking built on experience against requirements rewards the obvious resume match, and the person who does best in a job is usually the one you'd least expect. I once pushed a technology client to interview a CTO whose resume lacked direct industry alignment. That CTO helped the company raise its Series A and Series B, then co-founded new ventures with the original team.
How Can Employers Assess Candidate Judgment for AI Output Validation and Supervision?

Among Canadian CHROs, 68% named supervising, validating and overriding AI outputs as the workforce's most essential AI-era skill, while 29% of Canadian employees ranked judgement as important. Knowledge is a commodity now, so everyone has an equal opportunity to perform at a high level. What still separates people is hard work, creativity, a can-do attitude and instincts, and AI makes the average person better and the exceptional person that much better.
A person signing off at the end won't necessarily catch what the AI gets wrong. In a hiring experiment published in Frontiers in Psychology, about 60% of participants who got biased recommendations from an algorithm didn't notice the bias, even when asked about it directly. So supervising AI is a skill to hire and test for, including on your own recruiting team.
I ask candidates to walk me through how they thought about their workflow, how they applied AI and the thinking behind it. Then I give them a business problem and a goal and ask them to reverse engineer how they'd get there, and why. No single question settles it, so I ask in a few different ways and see whether the answers line up.
Whatever seat HR gets when your company defines its AI strategy, you still run your own recruiting workflow and hire the recruiters who'll supervise the AI inside it. I'd judge any AI you bring in on time to quality of hire, meaning how fast you land someone who actually performs in the role.