AI in hiring: more volume, less signal
AI has moved into hiring with force, and the effects are showing. Applications keep rising, hiring rates are slipping, and both sides feel stuck. The system is producing more noise than signal.
AI-led interviews and auto-written cover letters changed how people apply and how teams screen. The result: fuller pipelines, fewer clear fits, and slower time-to-fill.
What the data is telling HR
A recent survey from the Society for Human Resource Management found roughly a third of job seekers use ChatGPT during their search, and more than half of employers use some form of AI in recruiting. That mismatch creates volume without clarity. Employers see polished documents, but not necessarily stronger candidates. Source
Researchers studying tens of thousands of cover letters on Freelancer.com saw a shift after late 2022: letters got longer and cleaner, but relevance dropped. It's harder to separate true fit from confident fluff. As Jesse Silbert warned, if we don't improve how information moves between workers and firms, outcomes get worse, not better.
On interviews, Greenhouse reports that 54% of U.S. job seekers say they've had an AI-led interview. Bias doesn't disappear just because a model asks the questions. As Djurre Holtrop notes, algorithms can copy and even magnify human bias. Greenhouse CEO Daniel Chait describes a loop where employers automate and applicants automate, and everyone gets frustrated. Source
Why it's getting harder to pick the right person
- More applications, less differentiation. AI makes it easy to send "good-looking" materials at scale.
- Over-optimized resumes and letters hide actual skill. Writing quality goes up while signal goes down.
- Automated interviews add steps but don't always add insight. Bias risk stays if you don't use structured scoring.
- Hiring rates and starting salaries trend down when fit is unclear and selection confidence drops.
Practical fixes HR can ship now
Reduce noise at the top of the funnel
- Write precise job requirements. Spell out must-haves and nice-to-haves. Add knockout questions tied to real work.
- Close cover letters or replace them with 2-3 short prompts specific to the role (e.g., "Share one example that shows you can do X in our stack/process").
- Use early work samples: brief job simulations, realistic cases, or writing exercises using company-context. Randomize versions and set clear scoring rubrics.
- Limit open application windows or set a cap per role to keep review manageable.
Make AI help, not decide
- Use AI to summarize, cluster, and de-duplicate applications. Don't use it as the final selector.
- Adopt structured scoring with anchors for each competency. Blind irrelevant fields (name, address) to cut bias.
- Audit prompts and outputs on a schedule. Test for drift and disparate impact across groups. Keep review logs.
- Have a human reviewer sign off on every move past key gates.
Retool interviews for clarity and fairness
- Run structured interviews with the same validated questions for every candidate. Score each answer against clear criteria.
- If you use asynchronous or AI-assisted interviews, share instructions, context, time limits, and scoring guidelines up front. Include one practice item.
- Favor evidence over performance: portfolios, repos, case write-ups, and paid trial tasks beat clever small talk.
Protect candidate experience
- Set response SLAs for each stage and stick to them. Tell candidates what to expect and why.
- Keep take-home tasks short (60-90 minutes). Pay for anything longer.
- Close the loop with brief, useful feedback when possible.
Measure what matters
- Track pass-through rates by source and stage, time-to-fill, quality-of-hire, offer acceptance, and candidate satisfaction.
- Watch for sudden changes after adding tools or prompts. Investigate gaps across demographic groups.
Stay ahead of compliance
- States including California, Colorado, and Illinois are moving on rules for AI in hiring. Expect notice, audit, and transparency requirements.
- Document purpose, data used, retention, vendor responsibilities, and fairness testing. Offer alternatives when candidates opt out of AI-led steps.
What this means for HR
AI isn't going away. Use it to remove admin work and improve review quality, not to outsource judgment. The goal is simple: stronger signal with fewer steps.
If you want practical training for your team on ethical, effective AI in hiring, explore curated courses by job here: Complete AI Training.
Key takeaways
- More applications don't mean better hires. Raise the bar on information quality at each stage.
- Structure beats style. Use rubrics, work samples, and consistent interviews.
- Keep humans in the loop, audit your prompts, and track outcomes. Fix what the data exposes.
Your membership also unlocks:
