What to consider when incorporating AI into the talent acquisition process

AI has permeated so many aspects of work life, especially for HR leaders. But its presence in talent acquisition (TA) has proven particularly hairy, as professionals navigate questions around ethics, bias, and the importance of human touch.

Adam DeRose, HR Brew’s resident tech reporter, recently sat down with TA leaders at a From Day One conference in Washington, DC to gather perspectives on and analysis of the role of AI in modern TA.

Human touch. The prospect of losing human touchpoints is among the biggest concerns that many people have with AI coming into HR departments. Panelists addressed it head-on, noting that AI frees them up to have better interactions.

“What we are enabling our recruiters, from a capability, learning and development standpoint, is that these tools are here to automate the process, but not remove the humanity,” said Shabrina Davis, head of manager enablement and inclusive hiring, at Amazon. “It allows us to focus on the people part of talent acquisition, and then those processes that are fully automated really give us the chance to make sure that every candidate has the same experience.”

Meghan Rhatigan, VP of global talent acquisition experience at Marriott International, agreed. “What AI brings, is the opportunity for us to create more capacity for our recruiters, so that they can be more human-centric,” she told the audience in agreement. “So they can have the opportunity to have more time to have conversations with their candidates, have conversations with their hiring managers, build relationships, whereas before they were like, clicking, clicking, clicking, clicking, clicking, just to get somebody hired.”

When to automate. Obviously, not everything can be automated…yet. (Though we wish there was a way to automate cleaning the fridge, but we digress.) When considering what to automate, Rhatigan said Marriott looks for repetitive tasks. “There’s a saying, ‘If you do it twice, automate it,’” Rhatigan said. “For us, we’ve started to do a really intensive look at what are the transactional, repeatable steps that our recruiters are taking, our coordinators are taking, our hiring managers are taking, and those are the first places that we're looking at and already have looked at automating with AI.”

Automating interview scheduling, for example, has helped Marriott hire more efficiently. “Our executive recruiters and our coordinators are no longer scheduling interviews for executives. They’re fine. No one is complaining that they got a text message to schedule an interview,” Rhatigan said.

The bias factor. There’s been an ongoing debate over whether AI will reduce or introduce potential race, age, or gender biases into the hiring process. The HR leaders on stage said they take this seriously.

“There is undoubtedly bias in AI. There’s bias out there, because humans are biased. And the more that humans feed into the LLM models, that behavior is the data,” Rhatigan said. “I don’t care what anyone says, and maybe it’ll get better over time, but because of that, we’re [Marriott] just not going to put ourselves in a position where AI that’s ultimately bias is making decisions on behalf of our company.” She added that Marriott does not allow AI to make hiring decisions to reduce the chance of bias.

Davis had a different take on how AI can benefit hiring. “If it’s a strictly human experience, then that bias is there. And certainly any technology that you add to this process, there is room for that, but it also solves it, because it levels the playing field,” she said. David added that technology can identify areas of potential bias and help the company remove bias from the TA process. “Does it [the AI] create fairness? Does it create a leveling field for our recruiters to be humans in the process? Is the candidate experience improved? Or is there room for improvement?”

Jason Long, senior HRIS analyst at Globalization Partners, a firm that connects employers with international employees, said the fear of bias largely comes from skepticism around AI’s trustworthiness. “It’s been interesting to see that we don’t necessarily have a lot of trust in AI and what it can do,” he said. “I feel like we’re almost at, like, that dot com moment or something. When somebody figures out how to make it to where you can trust AI, I think it’ll be a big game changer.”