Select Page
pulse.

Remember when things just seemed to work? Yeah, neither do we.

Get our free email series 5 Days to Less Dysfunction below.

Is Your Algorithm a Racist?

by | Jul 7, 2016 | The Workplace

Using algorithms in the hiring process can open your business up to some unwanted consequences.

Iwas watching a conversation the other night about artificial intelligence, when the topic turned to using algorithms in predictions, particularly in the job hiring process. Though still in its infancy there are a number of “workforce solution” companies that sell access to such algorithms to businesses of all sizes. They promise to deliver hiring managers the best of the best, the most likely to become top performers, and do so while removing human judgement from the process.

Inherent Flaw

Now I being one who spends much of his time living in the world of human biases and decision-making, my ears always perk up when I hear such great promises made about removing human judgement. To be clear I am a big proponent of data in terms of providing benchmarks and guideposts, but there is an inherent flaw that exists when algorithms are touted as human replacements….algorithms are designed by humans.

Silicon Valley can be thought of as the Algorithms Capital of the World, but it also has been seen to be one of the most racially, ethnically, and genderally homogeneous places to work. Even companies like Google have acknowledged this over the past few years and have made efforts to become a more diverse company, recognizing the value such diversity brings and also the responsibility they have for making it so. The question is, how did such a forward-thinking place as Silicon Valley become so white and male? The answer in large part lies with a cognitive bias called In-Group Out-Group.

In-Group Out-Group Bias

During our early beginnings, forming groups was crucial to survival…to fend off attacks and to survive in harsh environments. Those who were not part of those groups, often didn’t survive, and as such they didn’t contribute to the evolutionary gene pool. Put another way, we are hard-wired to form groups.

Groups can form on minor, even trivial criteria. Nonetheless, they will find the other members more agreeable and identifiable than those outside the group, even if the underlying affiliation is as mundane as living in the same city or working for the same company.

On the flip side, people tend to view those outside their group to be more similar than they really are (“out-group homogeneity bias”), basically clumping all of the outsiders together and forming prejudices and stereotypes about them.

Now while this sounds very sinister, the fact is, often it’s happening at a sub-conscious level, and as a result, people take actions without realizing what was driving them to do so. This is where the issue of creating those job hiring algorithms comes in. Because the creators of the algorithms themselves are within a homogeneous group, the algorithms they create can reflect their biases. Parameters that on the surface look to be innocuous can potentially be excluding large swaths of potential applicants, and nobody would know why it is happening. Think a vicious cycle of chicken and egg.

Protect Yourself and Better Your Business

Remember what we said earlier about how these workforce companies promote the algorithms as an improvement over human judgement? Well therein lies the biggest problem…when you as a business owner or hiring manger purchase those algorithms you are shifting responsibility onto the math, but the ultimate hiring decision will still be on you.

So what should you do?

Here’s three things that you can do to benefit from such algorithms while also safeguarding your business:

1) ASK – You’re not going to get these private companies to show you their algorithm, it’s how they make money and they will cite intellectual property. Besides unless you are also a mathematician you wouldn’t know what you’re seeing anyway. So instead, ask to meet the designers of the algorithms. See for yourself how diverse a group they are. Also find out what specific safeguards they put into their algorithm to guard against biases. If they can’t specify any, or worse, seem dumbfounded at the suggestion that their equations would have any biases contained, drop them. They are operating from a position of either blindness to their own fallacies or are so overconfident that you will eventually find they have opened you up to possible problems.

2) REVIEW: Check the results of the applicant pool being sent to you for personal interviews…get a spreadsheet out or a sheet of paper and just track gender, race, ethnicity, and other background criteria. If you see that one group is over represented then that is a good indication of potential bias within the algorithm…and as such reverting back to Step 1 is a must.

3) BALANCE: Sure you want to hire the best performers for your business, that makes perfect sense. But, there is a balancing act between the best performers on an individual level, and the quality of your business as a whole. Does the need for an extra 1% of sales that a top performer will bring in worth more than being a workplace that embraces fairness and opportunity? Are you sure that somebody not deemed “best of the best” can’t be trained and mentored and developed into such a top employee? Isn’t that part of what business leaders do?

I’d like to think it is, and so do many people working hard to better themselves and whom are just asking for a chance to prove what they can do.

If only there was a math equation to reflect that.

 

Robert Westfall

Robert is a writer, behavorial researcher and decision-making consultant. He is the founder of Instinct, a firm specializing in helping organizations be more human focused and planet conscious.  You can learn more about his work at www.TheHumanInstinct.com and follow him at twitter.com/WeAreInstinct

RECENT WRITINGS

News Flash: Facebook Ain’t News

News Flash: Facebook Ain’t News

Facebook’s continuing desire to be a leading source for “news” is a sad indictment on the state of credibility and critical thinking in the world.

newsletter-signup-pulse-HOMEPAGE-V2-INLINE

Remember when things just seemed to work? Yeah, neither do we.

Get our free email series 5 Days to Less Dysfunction below.

RECENT WRITINGS

News Flash: Facebook Ain’t News

News Flash: Facebook Ain’t News

Facebook’s continuing desire to be a leading source for “news” is a sad indictment on the state of credibility and critical thinking in the world.

MORE WORKPLACE STORIES

Open up and say aaahhh…

Sometimes people are just where they are supposed to be, working with people in the right context and in the right manner. You as a manager need to recognize it and nurture it.

Hello old friend

As a small business scraping to get by amid tough competition often the best use of resources is doubling down on what you already have got.

Spread the knowledge love.

Pin It on Pinterest

Share This
Check your Pulse whenever you want delivered right to your email. 
x