As an HR professional, I’ve heard this a lot recently: “I know I’m qualified. I just don’t know if anyone — or anything — ever actually saw my resume.”
That question isn’t paranoia. It’s a reasonable response to how hiring works today…and the gatekeeper no one is paying attention to. AI-powered applicant tracking systems (ATS) have gone from a back-office efficiency tool to the default first gatekeeper in most hiring processes. And the leaders who run those processes are increasingly willing to say the quiet part out loud: the system is working exactly as designed, and it’s also screening out people who should be getting interviews.
The Numbers Don’t Lie
By 2025, 83 percent of companies were using AI to screen resumes, according to a Resume Builder survey of nearly 1,000 business leaders. That’s not a niche practice anymore, but the norm. And here’s the scary part: 88 percent of those same employers admit their ATS platforms are screening out highly qualified candidates, often because of formatting issues or missing keywords rather than any real gap in skill or experience.
Additionally, Resume.org found that 57 percent of companies already use AI somewhere in their hiring process, and 74 percent say it’s improved the quality of their hires. But more than half of those employers also voiced real concern about AI screening out qualified people or introducing bias. SHRM’s own analysis confirms it: 19 percent of organizations using AI or automation in hiring confirmed their tools had overlooked or screened out qualified applicants.
So, we have a system that companies believe is helping them hire better, that they also acknowledge is quietly rejecting good candidates, and that they’re expanding anyway. This is clearly a business risk hiding in plain sight.
Bias Doesn’t Disappear. It Just Gets Faster.
The promise of AI in hiring was that a machine wouldn’t care about a name, an age, or a resume gap. However, reality is far messier than expected. Employers themselves report where they see it happening: 47 percent identify age bias in their AI hiring tools, 44 percent point to socioeconomic bias, and 30 percent cite gender bias, according to research compiled by ResumeBuilder.
Academic research is confirming what those numbers suggest.
Stanford researchers reported in late 2025 that AI resume-screening tools rated older male candidates higher than equally qualified women and younger candidates — even when the resumes were built from identical underlying data. And the legal system is starting to treat this as more than a PR problem: a federal judge overseeing the Mobley v. Workday case ruled that AI screening tools can be considered an “agent” of the employer, which means the company, not just the software vendor, owns the outcome.
Regulators are catching up, too. New York City’s Local Law 144 already requires annual bias audits of automated hiring tools. California finalized its own AI hiring regulations in late 2025. Colorado’s AI Act takes effect in mid-2026 and will require companies to use reasonable care to prevent algorithmic discrimination. If you’re leading a hiring function and you haven’t asked your vendor for their most recent bias audit, that’s your next move — not someday, this quarter.
What This Means If You’re the One Hiring
Here’s my take, and I say this as someone who has spent a career helping leaders build teams that actually perform, efficiency and fairness are not competing priorities. They’re the same job done well.
If your AI screening tool is filtering out 1 in 5 qualified candidates — which is roughly what SHRM’s data suggests is happening — you are not saving time. You are paying recruiters to fish from a smaller, worse pond and calling it productivity. The average cost per hire already sits around $4,700, with a time-to-fill near 44 days. Every qualified candidate your system silently rejects makes that number worse, not better.
The fix isn’t abandoning AI. It’s supervising it like you would any new hire: with clear expectations, regular check-ins, and accountability when something goes wrong. Audit your tools and ask your vendors hard questions. Give a human a real chance to override the algorithm before a qualified candidate disappears from your pipeline.
What This Means If You’re the One Applying
If you’re on the other side of this, the strategy has to change with the system. Mirror the language in the jobposting. Use standard section headers. Spell out both the acronym and the full term for your skills and certifications. You’re not just writing for a hiring manager anymore. You’re writing for the machine that decides whether a hiring manager ever sees your name.
The talent is out there. The question every organization needs to answer honestly is whether their algorithm is actually finding it — or quietly deciding who never gets the chance.
