Friday, November 01, 2019

You no longer need a thumb to give the thumb’s down

I will be the first to admit that, when it came to hiring, I made a couple of mistakes that were real lulus. Once I brought an internal transfer onto my team, even though I knew he didn’t want the job and would be terrible at it. As it was put to me, I could hire this fellow or lose the headcount. Silly me. I took the guy. Net-negative.

There were other times when I hired folks even though their references were flashing WARNING signs. Big flashing WARNING signs. Nevertheless, I persisted in hiring them, assuming that my management genius would lay (figurative) hands on them and cure them of what was holding back their careers. What was holding back their careers was that they were all on the tail end of the nuts continuum. Alas…

On the other hand, my good hires were great. Exactly what I wanted in a direct report: good at their jobs, able to work independently, came to me when they needed help – but not for every little thing. In turn, for these types of folks, I was an excellent manager. I was hands off, supportive as needed, gave credit where due, and gave “my” people whatever level of visibility they wanted higher up in the organization. So, win-win.

Anyway, as a manager able to discern who was going to succeed and who was going to be a major royal PITA, I was a mixed bag. (That said, as a frequent interviewer for those who wouldn’t be reporting to me, but to would be directs to my colleagues or “superiors” [yuck: I hate that word], I’m certain that on occasion I was able to help someone dodge a hiring bullet.)

If I were hiring now – and the next hire I make will probably be for a home health aide – I might be able to take advantage of an “AI-driven assessment,” through which potential hires were ranked and received an “employability score.

HireVue is apparently the top dog here:

HireVue’s “AI-driven assessments” have become so pervasive in some industries, including hospitality and finance, that universities make special efforts to train students on how to look and speak for best results. More than 100 employers now use the system, including Hilton, Unilever and Goldman Sachs, and more than a million job seekers have been analyzed. (Source: WaPo)

HireVue has “proprietary technology that claims to differentiate between a productive worker and a worker who isn’t fit, based on their facial movements, their tone of voice, their mannerisms,”

There are plenty of doubters (including those from the AI world) that this approach works. These naysayers believe that using a system like HireVue is not just fraught with “pseudoscience,” but could be detrimental to “nonnative speakers, visibly nervous interviewees or anyone else who doesn’t fit the model for look and speech.”

Hmmm. I’m guessing that human interviewers might have an inherent bias against hiring “visibly nervous interviewees” or those who otherwise don’t “fit the model”. That said, a human interviewer may have a broader definition of who fits the model. Just say no to nosepickers, but overlook a quirk or two.

HireVue – not surprisingly – thinks its critics are wrong.

Loren Larsen, HireVue’s CTO argues that his system:

…is still more objective than the flawed metrics used by human recruiters, whose thinking he called the “ultimate black box.”

“People are rejected all the time based on how they look, their shoes, how they tucked in their shirts and how ‘hot’ they are,” he told The Washington Post. “Algorithms eliminate most of that in a way that hasn’t been possible before.”

One job-seeker interviewed by “the job-interview machine” had studied up on it in advance.

Nicolette Vartuli:

answered confidently and in the time allotted. She used positive keywords. She smiled, often and wide.

Despite best efforts, she didn’t get past the initial screen for the investment banking job she was looking for. And she was left wondering why she hadn’t been able to wow the interviewing bot. Was she too loud” Did she not use enough big words?

No way of knowing what the algorithms picked up on.

Of course, it may well have been that she only had a 3.5, and had gone to UConn rather than Yale. And not that she didn’t display enough enthusiasm, or that she only used the word ‘conglomerate’ once.

But with HireVue, job seekers don’t have access to their scores, or any feedback on what they aren’t doing right. (Humans can override the results of the first “interview” round.)

The HireVue system is attractive to “employers with lots of high-volume, entry-level openings.” While HireVue is the industry leader, there are competitors out there. One uses a bot that text messages candidates with some (dis)qualifying questions. Wonder if the folks being interviewed via text know that they’re communicating with a chatbot?

Even though a HireVue assessment may cover only a handful of questions, an interview “can yield up to 500,000 data points, all of which become ingredients in the person’s calculated score.”

Among those data points are “Facial Action Units” that measure “how excited someone seems about a certain task.” The system also measure word choice and tone of voice.

An individual’s data is measured against a benchmark of existing workers who hold a job. If someone measures up well against those who are succeeding in a certain job, they get a leg up on being hired.

Those who’ve bought (bot?) in claim that, with HireVue, time to hire speeds up, thousands of interview hours are saved, recruiting costs go down… Reading these results are giving me PTSD stemming from all the ROI studies I conducted over the years, all the metrics I got someone to agree to my estimates on…

Whenever I see a claim that “using your miracle product, we saved $1M bucks a year”, my first thought is “and you really have no way of proving that.” Just saying…

There are also non-metric “results.” Here’s what one HR person had to say about using an “AI-driven assessment.”

“The more digital we become, the more human we become.”

How so? (Can you quantify that???)

I have mixed emotions about this.

I’ve been rejected for plenty of jobs over the years, and maybe there would have been less of a sting if I knew that I’d gotten the thumbs down from a machine rather than from a human. But a machine tossing my resume because I didn’t pass a keyword scan is one thing. Having to “interview” with a machine is another. “Alienating” and “dehumanizing” are words that some interviewees used. I’m with them.

On Reddit, someone offered a tip for dealing with a HireVue interview:

“Glue some googly eyes to your webcam. It’ll make it easier to maintain eye contact.”

While I’m not 100% enthusiastic about growing old, it’s times like this when I’m happy to be well out of the career fray. Having to worry about gaming an AI “interviewer” by attaching googly eyes to my laptop….

Two thumbs down on that one. I’ve got the two thumbs to be able to do it, which could be seen by the webcam but not, I suppose, by the googly eyes.

Ah, progress!

No comments: