Concerning AI recruiting, I am not concerned
Concern over AI recruiting is misguided. Here's why.
Nearly everyone agrees that hiring is broken.
Candidates complain about ghosting, marathon interview processes, impossible job requirements and convoluted assessments. Hiring managers are overwhelmed with thousands of unqualified applicants, generic AI-crafed resumes, and juggling hiring alongside their full-time job.
If that’s not enough to get you down, apparently SkyNet AI is about to appear in full force and really tear sh*t up. Take this quote from a TIME magazine article on how bad job hunting is and how it’s about to get worse thanks to artificial intelligence:
“When you get a form rejection at 2 a.m. after you’ve applied for a job at midnight, Reid says, you can be pretty sure software is sending that message. As frustrating as long and onerous interview processes are, the increasing reliance on artificial intelligence, with its lack of nuance and context, could be even worse for job seekers.”
Then there’s this from the Harvard Business Review, which compares new hiring technology to the ancient practice of reading tea leaves:
“Then the process moves into the Wild West, where a new industry of vendors offer an astonishing array of smart-sounding tools that claim to predict who will be a good hire. They use voice recognition, body language, clues on social media, and especially machine learning algorithms—everything but tea leaves.”
And this article in the Wall Street Journal that compares modern recruitment to a game of Pac-Man where candidates must defeat the final boss:
It’s bot versus bot on the hiring front lines, as job hunters turn to new tools to counter the AI that’s screening their applications.
I get it: AI is yet another boogeyman in a long line of boogeymen for the very real frustrations that job seekers have. And it doesn’t hurt that new boogeymen are a pretty easy way to generate clicks.
But this line of thinking is problematic for two reasons: it deflects blame from companies with poor hiring practices and creates a stigma around the adoption of tools that might actually—hear me out here—help.
AI hysteria conveniently ignores the fact recruiting has been broken for a long time
Even if you’re ready to dismiss me as an AI-hawking huckster, we should be able to agree on the following:
For most open roles, when a candidate submits a resume, it is typically reviewed in about 30 seconds by a recruiter or hiring manager. 30 seconds is barely enough time to tie your shoelaces, let alone review a resume—regardless of how much we trust our resume-reviewing heuristics.
Even before the great AI-fication of everything, companies used basic knockout questions around work eligibility and experience, as well as ATS keyword screening to filter and rank resumes. There’s nothing new about being auto-rejected for a role.
The rise of globalization, remote work, and job boards with “Easy Apply” features have created an overwhelming number of unqualified candidates. AI is going to accelerate this problem as candidates create AI agents to spam companies with applications.
If you’ve hired a lot in your career, you’ve made some mistakes along the way. These experiences might have made you a competent judge of talent, but that doesn’t mean every single one of your peers is where you are now.
If we delve into the last point, the argument that AI filters are problematic implicitly suggests that human filters are always reliable. You can even start to see companies call this out in their job descriptions.
Yet there are studies that show recruiters are only a bit better than a coin flip at making value judgments on applications, and they still all disagreed with each other about what a good candidate looks like. The same study even found that relatively simple algorithms did a better job of assessing candidates.
“Despite the very limited types of data we input into both models, when presented with out-of-sample candidate profiles, both models made predictions more accurately than human recruiters.”
When I sent this article to a recruiter acquaintance of mine I got no response. I get it. The idea that something as innately human as judging others can be outsourced to a “robot” is unsettling. It seems almost dystopian to let algorithms decide our fates, but the alternative has not been particularly humane.
AI won’t fix recruiting, but it can help
In the near term, AI won’t make companies with broken hiring processes—those who scan simply on university name or who respond to applications with silence or who generally treat candidates like garbage—go away.
Part of the challenge with AI recruitment is that the hype is so great that its detractors have an easy time picking it apart. To me it’s a moot point, but for the sake of setting the record straight, I agree, AI won’t fix recruiting (at least until our machine gods begin recruiting each other 😅).
The idea that any technology in its current incarnation could fix something as complex as how the world hires is the stuff of sci-fi dreams; in fact, it’s a straw-man argument.
For example, I’ve spent the past decade doing marketing, and I can tell you with some authority, the advent of the internet didn’t solve marketing (i.e., remove the fundamental challenge of convincing people to hand you their hard-earned cash). However, it sure made it a whole lot easier to reach people you want to sell to, and, in turn, help those people find you.
Similarly, we believe that AI in the short term can:
Make it easier to surface the best candidates for open roles beyond simple knockout questions and keyword matching.
Make it possible to find diamonds in the rough; great folks who might get knocked out by today’s blunt force keyword screeners.
Reduce the initial evaluation/workload on recruiters and hiring managers so they can focus on having more in-depth late-stage interviews.
More objectively assess soft skills and the biases that humans bring to the table (like the firmness of a handshake).
Provide more immediate and consistent feedback to candidates, and nudge the hiring team to follow up. (Number one on the list of candidate complaints.)
It won’t solve the fundamental issue of human match-making, but it should help us reduce the noise and pain that it takes to get there. That’s an incredibly positive development and not something to be scared of.
The hype obfuscates the reality
In the end, hiring is a very human venture. Companies are nervous about not finding the right fit. Candidates are nervous about a huge life decision.
Any development that can make hiring more consistent, more comprehensive, more impartial, and more kind should be welcomed.