A 3 step guide on how AI can be HR leaders' best friend

October 10, 2023 3:19 PM

A person clicks on a smiley face with five stars underneath it.

If the last three years have proved anything, it's that the famous phrase, "With great power, comes great responsibility," doesn't just apply to superheroes, but arguably HR and people teams, too. Now, with AI transforming the workplace, those responsibilities have only grown.

Generative AI gives professionals the ability to categorize, pinpoint and summarize data and information in a matter of minutes. But like humans, technology is not infallible. A certain level of precaution and skills are necessary to effectively use AI, especially when it comes to hiring and reviewing employees, underlines Josh Merrill, co-founder and CEO of Confirm, a people management platform.

"One of the most common mistakes is to trust what the AI produces at face value," says Merrill. "When AI does answer a question, it does so in a very confident way that conveys a false sense of accuracy and authority to the user."

Confirm specifically helps companies minimize bias in their performance reviews and utilizes AI to summarize employee responses. In one case, Merrill recalls AI tech looking over 200 employee surveys and concluding that a small percentage of employees experienced harassment and discrimination. And yet, that wasn't actually the case.

"We went back to the original survey responses and read every single one, and there was no inference or suggestion of harassment or discrimination," says Merrill. "So we asked the AI, 'Why did you say that?' And the AI actually said to us, 'A small percentage of employees of any organization experience harassment and discrimination.'"

With the right approach and a dose of curiosity, Merrill is confident AI can help leaders better manage and evaluate their workforce, if users are willing to be thoughtful in how they interact with the tech. Here are the three best uses of AI in today's workplace strategies, according to Merrill.

Providing holistic feedback

While Merrill agrees that feedback should ultimately come from employees' managers, he notes that AI can ensure managers are providing constructive criticism as well as praise.

Read more:

What's the future of HR and AI? Compliance officers from HUB break it down

"Humans are pretty lousy at giving feedback, and a lot of managers really struggle with giving critical feedback," says Merrill. "The role AI can play is offering an outline of how that conversation should go and the talking points that might help you to have that conversation."

Essentially, AI can help structure notes on an employee into a conversation about their performance, taking into account company, team and employee objectives. Managers should check and revise the AI's work as needed, but the tech can at least lay the foundation for a productive conversation, says Merrill.

Identifying flight risks

AI can also help people teams pinpoint which employees are overworked and feeling apathetic towards their workplace. For example, Confirm uses organizational network analysis to conduct performance reviews, meaning anyone at a company can provide feedback to an employee. Merrill notes that employees will point out if someone seems to be working too hard or lacking support. From there, the AI can do what it does best: dissect the data.

Read more:

LinkedIn launches new AI tools to make recruiting easier

"It's pretty easy to have an AI read through pages and pages of documentation and then start to draw a conclusion," he says. "It can look holistically at an employee base, and say, 'This is the group of people who we think may be flight risks.'"

Reducing bias

AI has the power to present biased information as fact without much consequence. However, Merrill argues human managers are even more guilty, citing studies from 1998, 2000 and 2010 that all confirm the same point: 60% of an employee's rating will reflect the manager's bias, not the employee's performance.

Read more:

Don't invest in a return to office. Invest in better tech

"We're all talking about bias in AI, but let's also not forget how much bias there is in human beings," says Merrill. "But there's very little attention given to how biased humans actually are to begin with."

AI can at least be prompted to look for anomalies in ratings, so people teams and managers can take time to review how those employees are being rated. Merrill reiterates that AI is a tool whose answers should be consistently checked and adjusted — but it's also a tool that can pick out where humans are ultimately making mistakes.