AI and the Human Touch: Finding the Right Balance in Recruitment

The question isn’t whether to use AI in hiring anymore. Most teams already do in some form. The real question is where to stop.

Get it wrong in one direction and your recruiters drown in admin while candidates wait weeks for a reply. Get it wrong in the other and candidates feel judged by a black box, and the best ones walk.

Here’s where we’d draw the line.

Candidates want a human in the loop

The data is consistent. In a Pew Research Center survey, 71% of U.S. adults opposed AI making the final hiring decision, and 66% said they wouldn’t want to apply to an employer that uses AI to help make hiring decisions (Pew Research Center). In 2025, Gartner found only 26% of candidates trusted AI to evaluate them fairly (Gartner).

That doesn’t mean candidates hate technology. They like fast replies and easy scheduling. What they don’t like is a machine deciding their future with nobody accountable.

The rule: automate the admin, keep the judgment human

Let AI handle:

• Scheduling interviews and sending reminders
• Answering common questions about the role, the process and the timeline
• Sourcing lists and first-pass matching on skills
• Checking identity and enrolment
• Summarizing interview notes against your scoring guide
• Status updates, so nobody waits in silence

Keep people on:

• Final decisions. A named person should own every hire and every late-stage rejection.
• Interviews that judge motivation and potential. AI can help structure them. It shouldn’t run them.
• Selling the role. Students choose employers because of people: the recruiter who replied fast, the engineer who explained the work, the manager who seemed to care.
• Offers and negotiation. This is a relationship moment, not a workflow step.
• Late-stage rejections. Someone who reached the final round deserves a real message from a real person, ideally with one piece of useful feedback.

Be open about where AI is used

Transparency builds trust, and in some places it’s now the law. Since January 2026, Ontario requires employers with 25 or more employees to say in public job postings whether AI is used to screen, assess or select applicants (blogTO). New York City requires a bias audit of automated hiring tools and disclosure to candidates (NY State Comptroller).

Even where it isn’t required, tell candidates:

• Which steps use AI, and what it does
• That a person makes the final decision
• How to reach a person if something looks wrong

Watch out for automation bias

There’s a quieter risk: people trusting the machine too much. When a tool ranks a candidate low, it’s tempting to skip them without a look. When it ranks someone high, it’s tempting to wave them through.

Train recruiters to treat AI output as a suggestion with reasons, not an answer. Spot-check the candidates your tools rank low. Every so often, review a batch by hand and compare. If people and the tool often disagree, find out why.

Protect the human moments that win students

Students and new grads remember how a process made them feel. The moments that win them over are almost always human:

• A recruiter who replies within a day
• A personal note after an interview
• A chance to meet the team they’d actually work with
• Honest answers about the role, the pay and the path to grow

Use AI to buy time for these moments. That’s the whole point.

A quick self-check

Ask your team:

1. Could a qualified candidate be screened out by a tool without a person ever reviewing it? It should be no.
2. Do candidates know where AI is used in our process? It should be yes.
3. Does every finalist hear from a real person? It should be yes.
4. Do we test our tools for bias at least once a year? It should be yes.
5. Are recruiters spending the time they saved talking to candidates? It should be yes.