The influence of AI on our world is profound and ongoing, though its effects may be more understated than the sensational headlines suggest. Instead of the Matrix abound, AI is enhancing the work of human hands by simplifying or eliminating rote tasks, and making it easier for companies and workers to focus on more important tasks.

“AI is a job transformer, right? What it is basically doing is automating things, like high volume, repetitive tasks. And it is giving us more time to think and do something that we’re good at like problem solving,” said Ankur Saxena, SVP & head of strategic operations and talent at Mphasis.

One of those areas of work where AI is improving processes, and will continue to streamline on both the client and user end, is in hiring and talent acquisition. But there are many pitfalls, namely in how AI carries human biases. During a recent From Day One webinar, Matt Charney, talent acquisition leader at HR.com, spoke with professionals in-the-know about AI and how it affects talent acquisition.

The Genetics of Bias in AI

“There was one very famous article about Amazon creating an AI hiring bot, and it failed miserably because it was trained on data. The data was from all the people who are working in the firm. Being about 70-80% males, they unintentionally created a sexist AI hiring tool. It kept on selecting only people who are males,” said Saxena.

The matching of resumes to job descriptions has been a practice for over 20 years, with a historical feedback loop embedded in the machine learning and AI processes, says Dan Finnigan, CEO of Filtered.

To overcome bias Saxena says organizations need to look at the data AI is getting trained on, because the people training the AI carry their own inherent biases. Next you have to monitor the output so you can understand the results you’re getting. He likens this to going to the gym and maintaining an exact regimen but never seeing any increases and positive changes in your abilities. You have to change your workout to see different results.

Madeline Laurano co-founder of Aptitude Research turned the discussion to ethical AI and how it should be defined by transparency and specific use. “Ethical AI is pretty much defined by transparency. And are these providers going to be transparent with their algorithm? Or are they going to be transparent with the methodology that they’re using? Are they constantly re-evaluating it?”

Focusing on ethics, Charney directed the panel to think about two questions. First, will AI reduce bias going forward, or is automation bias simply replacing hiring bias? And within an organization, who is responsible for making ethical decisions behind AI-driven processes?

“I firmly believe that AI and recruiting is by definition biased, and maybe significantly so,” Finnigan said. He says that earlier in his career with Hot Jobs, product people found candidates using unseeable fonts to game the algorithm, and basically create a marketing document for themselves. On the company side, hiring managers would do the same, adding in things to make the job more appealing.

“And so it is biased by definition. It's just like the way we read news on social media; it's an echo chamber. So I would argue it's a bias accelerator so that we don't have to take the time to really try to figure out what's in the resume, or for the candidate to really figure out what's in the job description.” Finnigan says that the power of generative AI should be one that double checks bias and includes a process that is better at matching verified job skills, instead of just looking for patterns in applicants it's been trained to favor.

AI Is Still in its Infancy

Unfortunately, there aren’t a lot of companies out there that use AI very well for the hiring process, according to Laurano. She referenced Amazon's crash and burn with AI recruiting as a cautionary tale that’s still scary to a lot of talent acquisition leaders.

Charney turned to Alec White global head of talent acquisition at Computershare, which is in the early stages of that journey. White is working on the applicant tracking system. “We started with some fundamental things like digital interviewing, and self scheduling of interviews. And that, from the very beginning, felt natural. It’s the feedback loop that we’ve talked about.” White says that based on their metrics, their process doesn't “feel off putting to candidates, but like they are interacting with something human.”

“They could interview with us at midnight, with a digital interview, and then the system would tell them, ‘Hey, this is what is next’ and respond to questions and send them information that was customized to their role,” White continued. He says by personalizing the application process it doesn’t feel like a black hole with an automated email at the end saying not you.

Defining the Perfect AI Recruitment Tool

“I completely see it as an enhancer. I see AI as providing tremendous value to TA professionals, whether that’s being a campus recruiter all the way to a TA leader. “There’s lots of value in a lot of these use cases where AI can come in and improve the recruiter experience,” said Laurano.

She referenced research her firm did in early 2021 that looked at the recruiter experience. There were 14,000 job openings for recruiters that January, and the research found that they were wasting their time on tasks like managing job boards, manually advertising jobs, scheduling interviews, and more,  instead of connecting candidates with jobs. “AI can provide tremendous value in a lot of these use cases for recruiters. And I think recruiters that better understand and get excited about AI, they can get excited about generative AI."

“If you view a human as an algorithm, and you view AI as an algorithm, what do you trust as having less bias? We bring these biases into our organization, and it’s hard to unlearn those. But with AI, you can unlearn things, you can retrain it, and you can reduce bias in a way that you really can’t do with humans," Laurano added.

Editor's note: From Day One thanks our partner, Filtered, for sponsoring this webinar.

Matthew Koheler is a freelance journalist and licensed real estate agent based in Washington, DC. His work has appeared in Greater Greater Washington, The Washington Post, The Southwester, and Walking Cinema, among others.