Home Education AI in hiring: Speeding up recruitment, risking blind spots — why the human touch still wins

AI in hiring: Speeding up recruitment, risking blind spots — why the human touch still wins

0 comments

AI can’t see potential in a messy story or pick up on emotional intelligence. It can’t understand career transitions or sabbaticals or ask clarifying questions in a grey area. | Photo: iStock/ Getty Images

AI has made its way into nearly every stage of the hiring process, from resume parsing to scheduling interviews. Done right, it makes hiring faster. Done wrong, it makes hiring blinder. For Talent Acquisition leaders, the real challenge today isn’t whether to adopt AI. It’s how to balance automation with human judgment.

The problem with how AI is used in hiring today

According to Mercer Mettl (2024), 72% of recruiters say that while AI has increased hiring speed, it has actually lowered the quality of profiles reaching the interview stage.

Many things are going wrong. For instance, there is an over-reliance on keyword matching. Resume gaps are getting penalised without context. Biases are being baked into algorithms, especially around age, career breaks, or non-linear journeys.

AI isn’t neutral by default — it reflects the data it’s fed. If more women list parenting breaks, then AI filtering on continuous employment will disproportionately screen them out, even when they are equally or more qualified.

What AI is actually good at

I asked Vivek Pandey, Vice-President, Human Resources, Dream11 who shared – “While hiring, AI saves the initial screening and shortlisting time for junior level positions. However for mid and senior levels, it has to be clearly the ability of the talent acquisition teams to make an assessment. AI cannot replace the human element which is being sensitive to areas like work culture fitment, and traits like leadership, collaborative working and teamwork”.

Let’s not throw the baby out with the algorithm. AI is highly effective when used for: parsing structured data from thousands of resumes; creating job descriptions quickly and at scale; scheduling interviews and automating logistics; and flagging missing skills or formatting issues.

What AI should and should not be used for

What it should not be used for: final shortlisting; cultural-fit decisions; rejection decisions based on timelines alone; screening out returnees or career shifters. AI can scan, but it can’t sense.

AI can’t see potential in a messy story or pick up on emotional intelligence. It can’t understand career transitions or sabbaticals or ask clarifying questions in a grey area.

Good hiring is about contextual judgment. A resume doesn’t tell the full story. A conversation often does. Great TA teams know when to use data and when to trust their instinct.

How talent aquisition teams can stay relevant

Here are steps to modernise your hiring without compromising the human lens: Train recruiters in AI literacy; know how the filters work, what’s being excluded, and how to write better prompts.

Audit your AI stack quarterly. What’s being filtered? What’s being missed? Fix the blind spots. Create a Human+AI hybrid SOP, where AI assists the screening and recruiters make the decisions.

Candidates should know what’s automated and what’s human. Transparency creates trust and trust creates better candidate experiences.

Conclusion

AI can power your hiring engine, but humans still need to drive it. The future of recruitment doesn’t lie in choosing between technology or people. It lies in knowing when to let machines support and when to let people lead. Because in the end, great hiring has always been about one thing: understanding people.

Published – August 08, 2025 01:27 pm IST

You may also like

Leave a Comment

Welcome to Naija Global Television (NGTV) — Nigeria’s largest privately owned terrestrial television broadcaster, proudly serving audiences across 24 states. With our headquarters in Akwa Ibom State and a strong presence in Cross River…….

© 2025 Naija Global Television – All Rights Reserved. Designed and Developed by Pro