AI Hiring Tools Just Flunked Every Standard
Seven types of bias, zero proven validity, and a regulator closing in.
Inside this issue:
Why my firm just declared every AI hiring tool non-compliant with selection standards.
The four jobs AI cannot do in executive recruitment, no matter how good the next model gets.
A cardiovascular VP search where the strongest profile-match candidate did not get the offer.
Every week another LinkedIn post tells you AI is replacing recruiters. Every month another article warns executives their job search is obsolete. The framing is always the same: AI is coming for everything, including hiring, and the future belongs to candidates who can outsmart the algorithm.
It’s clickbait.
The Mercuri Urval Research Institute just published a report (Sjöberg, 2025) reviewing the AI-based tools companies are buying right now. The verdict: non-compliant with professional selection standards.
That isn’t a problem the next model release fixes. It’s a problem with the methodology itself.
Why the Clickbait Keeps Coming
Three groups profit from the AI-takes-over-recruitment narrative.
HR tech vendors sell screening tools to corporate buyers. The more the market believes AI is inevitable, the easier the sale.
Content creators chase engagement. “AI is coming for your job” outperforms “AI is mostly hype” on every platform, every time.
Career consultants pivot their offer to “AI-proof your search.” A panicked audience is a paying audience.
None of these groups operate inside an executive search firm. None of them have visibility into how senior placements actually happen.
I do. Here’s what’s really going on.
What the Research Actually Says
The MU RI report didn’t review a handful of tools. It reviewed the entire category. The findings should end the debate.
AI hasn’t beaten traditional methods
Across studies, AI-based selection tools produce gains that are modest, situation-specific, and outweighed by the risks they introduce. The research could not find meaningful improvement over validated traditional methods used in retained search for decades.
The methodology itself is broken
Proper selection has four steps: define the job, identify the required knowledge and skills, build standardized assessments, then combine the scores. AI tools skip the first three and automate only the fourth. They’re combining scores from data that was never validated against the actual role.
Bias is structural, not fixable
MU RI documents seven distinct types of algorithmic bias built into these tools. Representation, historical, measurement, aggregation, emergent, learning, and human. The last one is the killer: models trained on human ratings inherit every bias the raters had. The tools sold as “removing bias” amplify it.
Standards and regulation have closed in
The EU AI Act classifies personnel selection as a high-risk use case. Mandatory human oversight, documentation, transparency, and ongoing bias monitoring. Most vendor tools don’t meet that bar. SIOP, ITC, and ISO 10667 are the three bodies that set assessment standards globally. All three apply the same validity, reliability, and fairness requirements to AI as to traditional methods. AI fails all three.
This isn’t an early-adoption problem. This is a fundamental architecture problem.
The 4 Things AI Cannot Do in Executive Search
Reading vendor marketing, you’d think AI is one upgrade away from running searches end-to-end. Reading the actual research, the gap is structural.
Here’s what AI cannot do at the executive level. Not now. Not soon.
1. Read between the lines of a profile
A CFO’s LinkedIn says “led integration of three acquisitions.” AI matches the keyword. A human consultant calls a board member and learns the CFO opposed two of those acquisitions and was overruled. That context decides the placement. AI cannot access it.
2. Convince a passive candidate
80% of executive placements involve candidates who weren’t looking. The conversation that moves them is a human one: someone they trust, calling at the right moment, framing why this role is different. No automated outreach replicates that.
3. Run a confidential search
When a board replaces a CEO without public knowledge, the search firm carries the confidentiality. AI tools that log queries, train on data, and generate suggestions create surface area no general counsel signs off on at the executive level.
4. Be accountable for the outcome
A retained firm guarantees its placement. If the executive fails inside 12 months, the firm replaces them at no additional fee. AI vendors guarantee nothing. The accountability gap is the entire business model.
If you want this done properly, my LinkedIn Optimization service rebuilds your profile for LinkedIn Recruiter visibility.
Done by someone who runs those searches every day for executive mandates in life sciences.
What’s Actually Happening in Real Searches
AI is in the recruitment workflow. Just not where the clickbait says.
It speeds up sourcing for high-volume corporate roles. It helps consultants draft outreach faster. It surfaces resume data for human review. Useful as a tool. Not a replacement for the judgment, relationships, or accountability that drive an executive placement.
Last quarter I ran a search for a VP of Commercial Operations in cardiovascular devices. On paper, the strongest profiles were obvious: exact title match, same sub-sector, overlap with the right competitors. The candidate who got the offer at €295k didn’t fit that pattern. His most recent title was different. His sub-sector adjacent. He made the shortlist because of a 40-minute conversation that no profile match could replicate.
The fix wasn’t a better filter. The fix was a consultant who knew when to ignore the obvious matches.
That’s what every retained search looks like behind the scenes. Databases surface data. Humans decide.
What hasn’t changed for executives
The fundamentals of how senior placements happen are the same as they were five years ago. Be findable in LinkedIn Recruiter, where retained consultants run every search. Build direct relationships with consultants in your function and industry. Activate weak ties for warm introductions. Treat job portals as last resort, not first move.
None of this is obsolete. None of it is being automated away. The articles claiming otherwise are written by people with something to sell or engagement to drive.
The Real Path Forward
AI is a tool inside executive recruitment, not a replacement for it. The clickbait will keep coming because panic sells. The reality is more boring and more durable: humans run the process, and your job is to be visible to those humans.
Your Action Step: Unsubscribe from one AI-panic newsletter or LinkedIn creator this week. Replace the time you’d have spent reading doom content with 15 minutes optimizing your LinkedIn headline using a standardized job title, your industry, and your sub-sector. That’s what a retained consultant searches for. That’s what gets you found.
The best executive roles are filled the same way they’ve been filled for decades. By humans, in conversation, with judgment. Make sure you’re in those conversations.
Till next time,
Kristof
P.S. Founder tier gets you two things: a full LinkedIn profile review from me, and weekly office hours in the chat for any question on your search. The same lens I apply to retained search mandates, applied to your situation. Become a Founder →