What happens when an AI machine decides you are ‘easy going’?

By Shira Jeczmien

Updated May 17, 2020 at 08:47 AM

Reading time: 2 minutes

784

When it comes to scoring your personality from 0 to 10—decimal points included—it’s rather hard to feel as though anything is accurate or authentic. ‘Ability to learn’, sure. Is 10 out of 10 too arrogant? 6.7 too low? What about 7.8? Now imagine that it isn’t you scoring yourself, but rather an AI system scoring you according to your social media profiles, CV or even just your name, email or phone number. Welcome to DeepSense, a big data science company working with AI to predict things. Predict what exactly, you may ask? For example, how likely your potential employees are to maintain stability in the workplace.

I entered DeepSense as an ‘employer looking for a content writer’. I was then asked a few questions about the type of position I was looking to fill, and to select what score range I’d like this candidate to hit in fields such as ‘Attitude and Outlook’, ‘Stability Potential’ and ‘General Behaviour’, I was then asked to drop in a quick link to the potential candidate and let DeepSense do its magic. I of course added my own LinkedIn profile out of sheer intrigue to see what my own score would be, and whether indeed this AI thought I was suited for the job I anyway do.

Within a mere number of seconds and a spinning of a 3D globe icon, my score was revealed. And the results were bleak. DeepSense had reduced my entire career, education, and skills into numerical scores that quite frankly were worse than I’d imagined. 5.2 for ‘Stability Potential’, 5.8 for ‘Learning Ability’, 5.9 for ‘General behaviour’ (what does that even mean?), and 5 for ‘Need for Autonomy’. Overall my match score for the ‘Content Writer’ position was at a 59 percent match. Not ideal.

Screenshot-2019-01-31-at-10.00.08

Sure, I identify as slightly neurotic with a shorter temper than most, but I never looked at these traits as hindrances to my career nor my ability to deliver within a role. Having lived alone since the age of 18, endlessly hustling the creative industries and wiggling my way through back doors, founding my own company and managing a team of people, I thought that my ‘Stability Potential’ score would at least be hitting on an average. A 7, let’s say. But DeepSense was determined this was not the case.

At closer investigation into what DeepSense categorises as ‘General Behaviour’ I found that this means “The overall ability to get along with others despite personal challenges.” For ‘Stability Potential’, the AI company scores according to the person’s inclination “to give it their all before calling it quits”, and for ‘Need for Autonomy’ I found out that this means a person’s tendency “to work better independently”. What’s more is that DeepSense has summed me up in but a few sentences, writing “Shira can be friendly at times but does not hesitate from being critical when the situation so demands. Shira can be sceptical of what others have to say and can question them incessantly until Shira gets his/her answers”.

Screenshot-2019-01-31-at-09.59.52

Now, I understand the potential appeal in a tool such as this. It can, in theory, help recruiters eliminate their own personal biases while at the same time scan through thousands of applications that represent but a small number of candidates who actually match the desired role. In an interview with the Verge, co-creator of DeepSense Amarpreet Kalkat, says that personality is often what determines a good candidate and that the AI developed by DeepSense is here to conclude just that. When asked whether AI is fit to determine a person’s character, Kalkat answers, “From a relative point of view, how accurate is human judgement?”.

Whether or not AI systems will heavily enter into the recruitment field is no longer a question of if, or when. They are here and for the long run. The workforce is growing, jobs are becoming more specialised and with that, the need to vet through applications is largely understandable. The question instead needs to be whether or not AI should be used to replace human judgement of character and more importantly—if it is able to predict a person’s character as DeepSense claims to—is it ethical to use this prediction for or against a person? All I know is that it did not feel great nor accurate to be reduced to ‘Slightly Easygoing’ and ‘Slightly Sensitive’.

Keep On Reading

By Eliza Frost

What is the Gen Z stare, and why are millennials on TikTok so bothered by it?

By Charlie Sawyer

What is ketamine therapy, the psychiatric treatment healing famous Mormons Jen and Zac Affleck’s marriage?

By Charlie Sawyer

First look at $1 billion UK mini city where controversial HBO Harry Potter series will be filmed

By Abby Amoakuh

Tiktoker gets slammed by dermatologists for promoting dangerous caveman skincare regime

By Abby Amoakuh

Only at Coachella can you be caught saying the N-word and still perform without question

By Eliza Frost

The Summer I Turned Pretty is getting a movie. Could it be here in time for Christmas?

By Eliza Frost

Bad Bunny is not touring the US due to fear of ICE raids at concerts

By Eliza Frost

Glen Powell’s GQ photoshoot is a satiric look at modern day males—and he’s in on the joke 

By Charlie Sawyer

Who is Zohran Mamdani, the staunch socialist primed to become New York’s first Muslim mayor?

By Abby Amoakuh

Euphoria fans freak out as major storyline for season three gets leaked

By Payton Turkeltaub

Do Gen Z secretly hate their boyfriends? TikTok’s viral #IHateMyBF says yes

By Eliza Frost

Everything to know about Justin Lee Fisher, arrested at Travis Kelce’s home over Taylor Swift deposition papers from Justin Baldoni

By Charlie Sawyer

President Trump and JD Vance angry over the DNC setting up a taco truck outside RNC headquarters

By Charlie Sawyer

UK women who miscarry could face home and phone searches following new anti-abortion police guidance

By Fatou Ferraro Mboup

Celebrities call out Blue Origin for sending Katy Perry and Lauren Sánchez to space

By Eliza Frost

Is the princess treatment TikTok trend the bare minimum or a relationship red flag?

By Abby Amoakuh

Campaigners call for gamers who carry out virtual rape in the metaverse to be charged as real-life sex offenders

By Fatou Ferraro Mboup

Belgian court lets convicted rapist go free so he can become a gynaecologist

By Eliza Frost

Taylor Swift announces new album on Travis Kelce’s podcast. Everything we know about TS12 so far

By Eliza Frost

Sabrina Carpenter says you need to get out more if you think Man’s Best Friend artwork is controversial