Are AI Screening Tools Costing You Top Talent?

What are your thoughts on AI screening tools being used to filter through resumes?

Given we’re currently in a candidate supply heavy market (there are hundreds of applicants per position, no matter the role), it is no wonder that AI screening tools are increasingly being adopted, especially in terms of resume parsing, but there is significant cause for concern.

There are numerous cases where it is being misused and good candidates are being ignored. Yes, in this market employers can afford to be very specific about the experience they require, but many times we find that candidates with a broader skillset can be just as, if not, more qualified for the job.

However, misuse of AI screening tools by recruiters can often just pick out candidates who are the closest match to an arbitrary set of requirements. I remember there was a Harvard Business Review from 2021, which stated that 88% of employers agreed that qualified highly skilled candidates are vetted out of the process because they do not match the exact criteria in resume-parsing tools. 2021 was a candidate-led market, where they had the power over employers; if anything, things could be worse now.

There is also the argument regarding how AI could reduce bias in hiring; according to CPO of Greenhouse, Henry Tsai this is an advantage of resume parsing tools. But, anyone familiar with the “alignment problem” knows it’s not that simple. We’ve seen AI bias in the past, whether racial bias in algorithmic risk assessment tools like COMPAS or gender bias in resume-parsing. What steps are being taken to prevent AI screening tools from hypothesising a certain name/background would be better qualified for a particular role given existing data?

The Solution

So, what is the solution? Most employers and HR teams do not have the time to sift through CVs manually and I am not arguing that there should be no parsing tools at all. Certain dealbreakers (such as a candidate being based abroad for a role that does not provide visa sponsorship) should exist to lighten the load.

The only solution is to blend AI tools with human oversight. There are some good parsing tools out there like Plum.io, which don’t just follow arbitrary rules on hiring specific candidates, and look for other qualities. The key is for HR teams to use tools sparingly and only for certain necessary qualifications, not across the board. 

And of course, these tools are improving but as of right now hiring is still a people-driven process. Many times, you are hiring people you will be spending most of your life with. At Gigstart, for example, we make the most of personal relationships by trying to understand clients’ needs and candidates’ strengths, outside of binary requirements.

I’m not arguing not to use AI in hiring because with the current state of the market, that would be simply impractical. What I’m saying is run an audit. How is AI being used in your hiring process? Is there the potential it could be being misused? Then you could be missing out on top talent.