The Very Real Dangers of AI and How HR Tech Vendors Can Fight Them

By Maren Hogan:

We work with a LOT of Artificial Intelligence Recruiting plays and one of the things you realize as you do is that HR and Talent Acquisition are not welcoming these tools with open arms. In fact, many of them are actively against using AI within the recruiting and hiring process until some of the fears they’ve articulated can be addressed. One fear that marketers often cite is the fear that AI will cost recruiters their jobs.

But that’s actually NOT the number one reason HR Professionals are skeptical about AI in the workplace. There are a myriad (founded) of other fears they worry about. Here are the top fears and some ideas for vendors, practitioners and analysts on how to combat them.

Of course, #HRTech is exciting, but there are dangers in #ArtificialIntelligence. @MarenHogan gives her insight on the top fears and how vendors can combat them: Click To Tweet

Bias in the Algorithm

Wait, what? Wasn’t AI supposed to remove bias from the process by ensuring our human subjectivity was a moot point? Well, in a perfect world yes, but who creates the algorithm? WE DO! That means all our preconceived notions about gender, race, socioeconomic status and non-traditional candidates can totally make it into the batter. And even one rotten egg can ruin a cake.

How to Fight This: Obviously humans are not going anywhere, anytime soon. In recruiting anyway, adopting a set of standards that are applied to tools across the board is one idea. Another is to use the purchase or consideration of an AI tool as a time to reevaluate…your evaluation practices.

Vendors? Partner with a bias-busting service or platform to help your new clients not become the Skynet of the future. You could also help lead the charge with ethical coding standards and a code of conduct for programmers and employers using AI. (cough, ATAP, cough)

Candidate FOMO

In a tight labor market, candidate FOMO is a thing. The very same candidate you may have shortlisted but not actually chosen could be the candidate today, that you would DIE to have. Non-traditional candidates are a thing:

The precision of AI could mean that degree requirements, felonies, drug tests, and specific skills tests weed out those who could be, with a little spit and polish, GREAT for the role. And in this market…who wants to miss out on that?

How to Fight This: Again, revisit your job requirements. Some are just plain dumb. A four-year degree is needed for very few white-collar jobs these days that don’t need additional job-specific training. So why put it there? If you have an AI or automation assessment filter, why the heck are you worried about unqualified folks applying? Get to the meat of the thing and get rid of conviction histories, drug tests, education requirements or the typical 5-7 years of experience, especially if you evaluate your staff and find that none of those THINGS MATTER. I have worked with ONE person over the last few years who didn’t have to cut out something out of his job requirements and that person was recruiting EYE SURGEONS. Are you? K then.

Vendors? You don’t have to be just salespeople. At this stage in the game, the AI recruitment platform sale is still consultative anyway, so use this period to help your clients create an audit to forever transform the way they hire.

Same Old, Same Old

You guys, we kind of touched on this before but it bears repeating and this is different than plain old bias. The way we did things before cannot be the way we do things now. When you grab yourself a shiny new tool, it’s easy to just plug it into your processes (not easy, but a natural inclination) but that’s NOT what the goal should be. It’s time to stop using the same hiring practices we’ve always used because we’ve always used them. For many smaller companies, this is a no-brainer, but the longer you’re around and the more hiring you have to do and the more quickly you have to do it…well desperate people forget.

How to Fight This: Past practice is the bane of change management gurus. They want to help you change but it’s near impossible in some of today’s large companies. Is it possible? Yes. Is it likely? No. So you find yourself a champion at the top to integrate this into your hiring and then use that champion to make real change in the “way you’ve always done things.”

Vendors? I know integration into other, more commonly used systems is the goal, but if you look at what I advise for practitioners above, you’ll see that assistance in winning over the changemakers is your mandate now. Focus on creating content and processes that make it dead simple to make the technology change, as your clients will have to make the people and process changes.

Guess what: your shiny new #HRTech could have biased #ArtificialIntelligence. Find out what can be done to combat this in @RedBranch's latest blog: Click To Tweet

It’s a Crutch, Maren!

Discussing AI without discussing what it means for the critically thinking mind of the recruiter, hiring manager, or HR Professional is a fool’s errand, as in, it shouldn’t be run. It makes no sense to use AI if your own recruitment team doesn’t understand the basics. One big vendor marketing idea (and it’s not a bad one, I’ve used it myself) is that AI frees up recruiters and even admins to do the more strategic stuff. But if they don’t understand the basic activities of recruiting, how can they emulate them? Or to put it a different way, if every candidate is being screened, assessed and chatted to by a chatbot that then stacks and ranks them, what’s the point of a recruiter? Good point Maren!

How to Fight This: Well, this is on you, I’m afraid. There are lots of great options for recruiters and sourcers who want to be better at their craft. Social Talent and SourceCon Academy to name two off the top of my head. ATAP and HROS are also compiling resources, creating community and building standards for recruiters, talent acquisition pros and hiring managers to draw from, so there’s really no excuse to NOT invest in becoming the kind of recruiter who invests in AI to further their knowledge rather than in the hopes of not having to accrue any knowledge at all.

Vendors? You can support learning communities like this and create spaces for these often volunteer opportunities to flourish. Support a webinar, put on an online conference, offer to co-present with a customer who’s doing it right. There are lots of ways to support excellence in recruiting and sourcing.

Privacy and Control

Lots of jobseekers who are interacting with chatbots today have no idea their information is being recorded as though they applied for the same position in another way. While it may be differentiated in the backend of these systems, it begs the question:

What obligation are we under to disclose the data we collect via automated and artificially intelligent means to jobseekers?

It’s a super valid question and one that GDPR (and the conversations surrounding it) went SOME way in answering, but I think we know that we’re still not in the clear. For instance, some algorithms claim to predict candidates’ skills, intelligence level and personality by using Machine Learning to scan Facebook or LinkedIn profiles. Is this under the jobseekers control? What if it’s a passive candidate?

How to Fight This: Derp, don’t use those tools unless a) the person is actively applying for a job at your company and b) you’ve informed them you are using such tools. Feel like that will adversely affect applications? Then don’t use them at all.

Vendors? Is this your only play? If so, consider whether you’re solving a problem in recruiting or creating one. If it’s the former, make your case in the sales literature (and educate your clients on the proper use of a product that could infringe on the privacy rights of a candidate). If the latter…Y THO?

AI-enabled recruiting tools don’t have to be the end of recruiting excellence, a privacy nightmare, or cause us to miss the best candidates we didn’t even know we needed. It can be a boon to doing our jobs better, faster and providing candidates with the information and communication they’ve been asking for since time immemorial. You just have to be aware of the hurdles and then plan for them, both vendors and practitioners.