In the 1980s, Hans Moravec and other AI researchers discovered a curious finding that we now know as Moravec’s Paradox. Essentially, AI can very easily learn to do things that are “hard” for humans, such as advanced statistics and analysis. On the other hand, things that are very simple for humans, such as identifying colors or recognizing faces, can be incredibly difficult for computers to do.


In short: computers, algorithms, and AI are great at some things, but humans are better at others. So, how can we find the right balance?


It’s not about us vs. them

The concept that humans and machines have different strengths played out in research from Harvard. In the study, a breast cancer detection algorithm was able to detect cancer cells 92% of the time. However, the doctors were able to identify cancer cells 96% of the time.


This clearly shows that humans are better, right? But wait—the next finding was perhaps the most telling part of the study. By combining the algorithm with human experience and intuition, the team as able to identify more than 99% of cancer cells. This blending of strengths points to the incredible opportunity that is facing many professions, including those in talent acquisition.


For many business leaders, the question is often “either/or,” as in “I either need recruiters or I need an algorithm.” The truth is, that’s the wrong question to ask. The right answer is an “and” solution, not an “either/or.” We need to look at how to bring the best of algorithms to the table and the best that humans have to offer—that’s when we’ll arrive at the best solutions.


Balancing human and machine in recruiting

Now that you’re looking for “and” solutions, it’s time to think more practically about where that best fits into your recruiting activities. Our research shows the best place for inserting machine-based support is at the earliest stages of the operation. Here’s an example of what happens when that automation goes too far into the process:


A few years ago, SHRM’s HR Magazine featured a story about Amazon. The company had put an algorithm in place because it was struggling to hire technical talent fast enough—a challenge many employers can identify with. The algorithm was designed to review resume submissions, offer assessments to qualified candidates, and generate offers on the spot for those that successfully completed the assessment. In the process, candidates never once spoke with a recruiter or interacted with a hiring manager.


When I present on the impact of AI on recruiting and tell this story from the stage, it is often accompanied by an audible gasp from the audience, because this inherently scares us to think that an algorithm might be making hiring decisions for us. The consequences of too much automation in this case could lead to poor fit hires, and most companies don’t want to deal with those repercussions.


What we’ve found in our research is a sliding scale for how comfortable people are with algorithmic decision making. Algorithms can easily and unobtrusively be inserted into the hiring process in the early steps of screening, matching, and scheduling (Clara’s specialty). However, once we get closer to the hiring decision, we want more human intuition, creativity, and other insights in the mix.

Your turn

As you think about your own hiring process and your talent acquisition team, consider where that point might be for your own firm. Where is your comfort level, and how can you take advantage of the best that machines have to offer so you can improve the capabilities of your recruiting team?


Ben Eubanks is the Principal Analyst at Lighthouse Research and the author of Artificial Intelligence for HR: Build and Develop a Successful Workforce. He also hosts the We’re Only Human podcast, focusing on the intersection of people and technology in the workplace.