Artificial Intelligence and Workplace Psychometric Tests
They say there are three things you shouldn’t talk about on a first date – religion, politics, and sex. These are rules we strictly adhere to at Nanalyze because every article we write is kind of like a first date for first-time readers. More importantly, our regular readers expect us to stay on point. However we decided to make the rare exception when we saw an article on Tech Crunch a few weeks ago about how AI can now detect gay people with an average accuracy of 87% based on facial analysis like this:
(Our MBAs came up with a test that doesn’t use AI and has better accuracy at 90%. It just identifies everyone as straight.) On the heels of these “accomplishments” in detecting sexual preference, we thought it would only be a matter of time before AI can detect political leanings, but the truth is that marketing people hardly need AI for that.
With a nation that is so spectacularly divided, and with people vomiting their political opinions all over any social media channel they can get their hands on, it’s pretty easy for marketers to figure out which way everyone leans. We’re talking of course about Britain, because in our great nation of America, we stand together stronger than ever, ready to face the day when the tool we’re currently using to identify gay people with, will (credible sources say) wipe us off the face of the earth. The truth is that while detecting religion, politics, and sexuality are easy enough without AI, what we could use some help with are getting rid of all the incompetent morons we have to work with every day by not hiring them in the first place. In particular, we’re talking about psychometric workplace tests.
If you’re unfamiliar with psychometric tests, then the best way to describe them would be similar to an IQ test except that the output comes in a variety of different forms with a focus on things like aptitude and personality. Some like the DISC assessment will classify you as one of four colors (red, green, blue, yellow), with each color having certain attributes associated with it as seen below:
For example, “reds” are more likely to try and control a group or butt heads with other reds.
Perhaps the most popular psychometric test for workplace settings is the Myers Briggs test. If you have never taken this test, you should. You’ll have to answer a set of about 64 multiple choice questions before you’re given a 4-letter result out of 16 possible combinations (the writer of this article is an INTJ for example). You can then read various descriptions of how your four letters ought to behave, and you’ll find yourself relating to all the positive attributes and none of the negative. Here’s a really cool breakdown from Wikipedia on the various types:
In an academic setting, the Myers Briggs test can be useful to study how various “types” behave in groups. In the workplace though, psychometric tests are not so welcomed.
We recently spoke with a “Head of Talent Management” at a multi-billion financial services firm about putting Myers Briggs test results in the employee directory and he looked at us like we asked him if we could all take turns sleeping with his wife. Using psychometric tests in the workplace presents far too much of a litigious risk and no employer would even think about opening up that can of worms. We can just imagine seeing headlines like “Class action lawsuit filed by ESTPs at Google claiming they were passed up for promotion by INTJs“. Since that sounds like a nightmare waiting to happen, why not just test employees before they begin employment?
While using psychometric screening for candidates sounds good on paper, it’s not such an easy sell. One startup we came across which was trying to break the mold when it comes to employment screening was a now defunct startup called Cream.HR. This startup claimed that they had achieved superior explanatory power over Myers Briggs by using something called the “Big Five Personality Traits” which are seen below:
The tool used profit as a proxy for productivity, and claimed that hiring managers could increase the probability of hiring an above-average employee from 50% to 80%. However when it came to actually selling the tool, things weren’t so clear cut. One of the founders gave a post-mortem on why the company didn’t succeed and imparted some interesting takeaways as follows:
- Just because a test has “self-evident economic value”, that doesn’t mean you can sell it
- “Productivity” is very difficult to measure, especially as the size of organizations increase
- As tests increase in accuracy, employers become less likely to purchase them because they don’t want to hurt people’s feelings (i.e. get sued)
- The problem with tests that work is most people don’t do well on them – like the HR people who want to take the test before using it for hiring
Some other problems that Cream.HR encountered were even more ludicrous. One HR department wouldn’t pay the premium fees for the test (it was much more expensive than the Myers Briggs alternatives) because that fell outside their allocated budget for screening technologies, and the increase in revenues would have been attributed to some other internal group. They wouldn’t be rewarded for these revenues, but instead they would be punished for exceeding their budget for selection technology tools (rolls eyes). Long story short, it’s not as easy as you might think to sell a tool that’s not a gimmick and produces detailed reports like this one:
In the words of Cream.HR:
People want an accurate test that doesn’t discriminate against anyone that makes everyone feel good that’s dirt cheap and doesn’t take any time at all to take.
Probably not going to happen anytime soon, but the truth is that AI will soon allow us to employ psychometric tests like never before.
A traditional Myers Briggs test uses a fixed set of 64 questions to determine your temperament. Artificial intelligence could potentially pull from a database containing a limitless number of questions. It could also take into account additional factors like, a delay in responding to a question, your mouse hovering over a particular answer before selecting another, or even cross-check what you answered against all that crap you vomit all over social media every day. In fact, AI could even start to come up with questions dynamically to test you in a way that no recruiter could ever dream of. Add some IQ assessments and throw in some technical questions, and we may soon wonder why humans were ever allowed anywhere near the hiring process.
We’ve written a number of articles about how AI stands to replace recruiters, and we covered startups like PredictiveHire which uses short questionnaires to screen candidates based on custom KPIs. In that same article, we also talked about TalentPitch which uses things like real-life judgment tests, personality testing, intelligence testing, and language testing. In our article on 7 Artificial Intelligence Startups in Recruiting, we touched on a startup called HireVue which uses raw audio, text from speech and micro-expressions as input to determine competency.
Aside from the startups we’ve discussed before, we’re not aware of any startup that’s working on a workplace psychometric test that isn’t based on Myers Briggs, and that can be used as either a pre-employment screening tool or to analyze “optimal placement” internally for a firm where psychometric attributes could play a role in success. Most importantly, the tool would be optimized to use artificial intelligence so that it actually gets better over time. Any MBAs out there working on a workplace psychometric test startup with some legitimate science behind it (ideally from one or more experienced and well-published psychologists on your team) should drop us a line and we’ll get you into a future article. We’d love to hear more about what you’re doing and how you plan to circumvent some of the problems Cream.HR encountered.
Of course this all becomes obsolete when we start using people’s genetic information to screen for employment. It’s been proven that genetics is tied to intelligence (and consequently competency), something that makes many people uncomfortable (except for the Israelis.) Of course those of you who are paying close enough attention will already know that all the jobs are going to the robots anyways, so this whole thing becomes a moot point. Fresh college graduates are going to end up doing what they did before graduating – doing drugs and playing video games.
Pure-play disruptive tech stocks are not only hard to find, but investing in them is risky business. That's why we created “The Nanalyze Disruptive Tech Portfolio Report,” which lists 20 disruptive tech stocks we love so much we’ve invested in them ourselves. Find out which tech stocks we love, like, and avoid in this special report, now available for all Nanalyze Premium annual subscribers.