We need more humanity in hiring. Is it possible for AI to help?

Let me say, up front, that I am extremely, extremely skeptical about the use of AI in hiring.

I don't like it when HR whines about how many applicants it gets, and uses that as an excuse to shop out their initial evaluation process to an ATS (applicant tracking system). Especially since the internet is rife with horror stories about ATSs that freeze, that bug out and leave you with nothing to show after spending an hour trying to fill out an application, that filter out qualified applicants, that are just bad. I look at that trade-off -- saving maybe some amount of up-front human time looking at applications, vs the false negatives and hit to the employer brand that comes from a bad ATS -- and conclude that I wouldn't want to work for any company that made it. 

For clarity, I don't have a problem with anyone using an ATS. I used Greenhouse at Vox and really liked how much faster it made the process of evaluating candidates. What I'm railing against is when orgs set up their ATS to automatically reject people who don't meet specific qualifications. Usually, this comes down to easy-to-quantify things; college degree (whether or not it's really necessary), years of experience (which are often a poor predictor of performance), or just not hitting the right keywords

If you want to set up your attorney job opening to reject folks who don't have a JD, I can see that. But beyond actual license and professional degree requirements, I get really uncomfortable with it. 

So my skin started itching when I read The Economist's (fairly) new SPECIAL REPORT on how AI can improve hiring and Human Resource functions. 

Now, I'll start with what I liked. And I did like some things. As much as I feel that workplace norms and processes in general need to be more humanized (and I feel that very strongly), it's also undeniable that humans are susceptible to all kinds of biases that are persistently difficult to train out of ourselves, and darn near impossible to remove from others by way of training. As the piece mentions, people tend to hire people like themselves, in a lot of ways. Technology gives some really great opportunities for blind evaluation and testing that can help us evaluate folks on more even terms. 

I am wildly skeptical of the actual example they use for how tech can overcome bias, which is a startup that gives applicants games to play as part of the application process. Supposedly, they use these games to rate applicants on 50+ different traits using "proven neuroscience." I don't think it's impossible that it actually reduces bias in the hiring process, I just think that evaluating people using algorithms is used as an excuse to avoid talking to real, individual people about what they want in a job, what they want from a workplace, and what they can do.

If AI can successfully complement human discussion and evaluation, as opposed to replacing it, then I think we're getting somewhere. The example in the Economist article of the company using HiredScore (which I'm also wildly skeptical of) to automatically pull possibly well-qualified applicants from your own database of folks who have applied before would be great. Such a system could reduce the burden of the applicant to always be applying for jobs and shift some of that to hiring managers, allowing them to think about people who applied before and contact them if they think they're a match (which hiring managers should be doing anyway, but heck, if this makes that easier then that's a win). 

Disappointingly, there's nothing at all detailed about thing most promising words written in the Economist piece, attributed to Athena Karp (dope name) of HiredScore: "[job applicants] are rarely told why they were not hired, nor are they pointed to more suitable jobs. Technology is helping “give respect back to candidates."

I am 100% in for a system that encourages employers to give substantive feedback to folks it declines to talk to or hire. I am also extremely down for one that will refer you to other jobs you might be well-qualified for when you're rejected (especially if those are referrals to different companies). But other than the sentence I copy/pasted above, there's nothing about an AI that does either. It's implied HiredScore does, but I couldn't find anything immediate along those lines on their website. 

There's a lot to hate in this piece, though. Nvidia is praised by using AI to figure out that people who submit especially long resumes tend to do less well there, so they're graded down by the system. Even if that stat is true, which to be clear I 100% believe, its a dangerous thing to filter people out by because, for the vast majority of jobs, it's completely irrelevant. Even if 75% of applicants who submit too-long resumes are trash, you're still deciding, up front, that you're willing to throw out those other 25% without even looking at them, for the sake of saving yourself some time. All because, for some unknown reason, there may be a correlation (NOT causation) between people who don't thrive at your company and people who don't know, for whatever reason, that they should shorten their resume. 

Don't even get me started on HireVue, aka the source of every positive article you can find on the internet about video interviewing. I am actually way more open to the possibilities of video screening than a lot of people, but my skin crawls a bit about the idea of using video interviewing to reduce bias. Using an AI to interpret the body language in applicant videos is inviting so many things into your hiring process that aren't good for applicants, and aren't really what you're looking for, either (who hasn't had a really good manager who slouched and mumbled? Who hasn't known a dynamic public speaker who gets nervous in one-on-one interviews?). 

In short, this is all a problem of capitalism. I'm being completely serious. Part of why video hiring is catching on is because HireVue puts a lot of effort into PR. They sponsor conferences, they send speakers. ATS companies, hiring startups, and HR consultancies are putting out a lot of the professional development opportunities that exist for hiring managers and HR folks, especially the free ones (like webinars). Being part of the conversation is how you get ideas out there, and the existence (and biases!) of venture capital has a big impact on which voices are being heard louder than others. Fully five of the people that The Economist quotes in its SPECIAL REPORT are the CEOs or other representatives of the companies being discussed (I'm happy to see clients represented as well, though).

Given both the technological promise and the VC sexiness of AI, that effect is gonna be amplified in the hiring and HR space. "Be a good manager and treat people like people" isn't gonna have a logo projected on a big screen at an HR conference, after all. Plenty of smart folks in the field advocate for good, basic practices but those aren't typically the people putting free webinars about improving your hiring process.

People pay for marketing because it works. But I urge HR pros to remember, as our north star, that we're in a fundamentally human endeavor. Technology is rarely going to provide us with better answers to questions like "why aren't my employees happy?" or "who is the right candidate for this job?" than just talking to other people will.