The racial implications of AI and UX
Racism Can Be (Un)Programmed
The Racial Divide in Product Design
Back in 2009, HP released a webcam feature that tracked a person’s face during a video chat. Here’s the catch, the technology was accused of being racist because it didn’t recognize Black faces. A headline from Gizmodo at the time read, “HP Face-Tracking Webcams Don’t Recognize Black People”.
Fast-forward to 2019 when a Vox headline stated, “Yes, artificial intelligence can be racist”. You’d think the lessons from a decade prior would’ve informed the product designers of recent technologies customized to interact with humans but, sadly, racial bias has yet again found its way into cutting-edge technology.
This time around, the “machine vision” of the AI in self-driving cars seemed to have a problem seeing Black pedestrians. Research conducted by a team from Georgia Tech found that there is a “predictive inequity in detecting pedestrians of different skin tones in object detection systems.” In other words, pedestrians with darker skin (and people that are shorter than average, apparently) can be more predictively at risk of getting hit by a self-driving car than someone with fairer skin and average height. Studies already suggest that Black people are more exposed to danger when crossing the street, so is this really surprising?
Good Intentions Are The Road To…
This is not some sinister plot, at least let’s hope not. A computer is only as smart (or obtuse) as its programmer. The root of this issue is that the majority of the data used in technological advancements like facial recognition and machine vision are based on a certain “standard”. Much like the social standard of modern Western society, the data being used is based on predominately white men. If a machine is programmed to learn like a white man and see like a white man, the margin of error for anyone that is not a white man may be of critical importance. This issue is nothing new.
“Self-driving cars are by no means the first technology to fail when confronted by other ethnicities: Google’s image-recognition system notoriously failed to discern black people from gorillas. Almost every product design has failed to grapple with the reality of humanity, from Kodak color film that reduced dark skin to a pitch-black smudge; to motion-activated taps and driers that refuse to acknowledge the presence of a brown hand but will trigger for a white one.”
From “The racism of technology — and why driverless cars could be the most dangerous example yet.”
Alex Hern
March 13, 2019
The Guardian
The Role of Implicit Bias
The problem product designers and developers are running into time and again stems from implicit bias. This type of bias is that subconscious set of values or beliefs that are accepted without any critical thought or even when evidence to the contrary is presented. It is the shadow of explicit racism and most white people refuse to believe that they could be carrying racist ideas somewhere in the back of their brains. When implicit bias finds its way into data that is fed to AI for purposes like machine learning, the results can carry real consequences.
Data isn’t the only thing dominated by the white imagination. Boardrooms are still significantly comprised of white men. When the majority of tech companies are owned and led by white men, the norm will be to continue to focus on development around a white male persona. Only when the larger public has a chance to view emerging tech through a multicultural lens do the problems caused by implicit bias become evident. It’s a sort of a perpetual blind spot for white people that has to be pointed out as a problem over and over again.
How Can We Solve This?
So what is the solution? Earlier we observed that tech is only as good as it is programmed. Herein lies a possible fix. If founders, programmers, and board members weren't such a homogenous group, the data we would be feeding AI would be more reflective of actual humans and all our shades and sizes, and not the white-washed data that has been used as the “standard” model for decades.
It’s not all bad news! Programming AI to find and flag racist behavior is another powerful application of the technology that is designed to have the opposite effect as implicit bias. For example, Fama Technologies is using AI to screen job applicants for sexist and racist comments. The impact is real because over a tenth of job seekers had their profiles flagged for their bigoted sentiment. In this case, AI provides a solution to prevent biased people from attaining positions where they may discriminate. It’s not quite Minority Report - but it’s a start.
The Role of UX Design
The long term solution requires more Black people and POC to be involved in the user experience (UX) design of products. UX is currently a very in-demand discipline within the design field and it touches every industry, including tech. The demand for UX Designers is tremendous and much of the research that goes into it doesn't require programming or coding backgrounds. If we can’t join the boardrooms right away, we should at least get involved with designing experiences that won’t put our lives in jeopardy.
I recently completed my first UX wireframe and all it took was studying some terminology, conducting a few surveys and interviews, and a little graphic design. As the process goes on there is more to it but it is not hard to start. In 3 years or so with more experience, a job in UX Design could bring in an income upward of $90,000. It doesn’t come overnight but young people that identify as non-white should seriously consider learning about UX design, especially digital natives interested in marketing.
The Best Way Forward
This is not a takeover and the objective isn’t running off white people. However, we need a seat at the table if we want to avoid the dangerous implications of artificial intelligence being programmed with implicit bias. Being compared to gorillas by Google images or run over by an autonomous vehicle because it might not discern us from the road isn’t cool. Our input is valid and the opportunity is ripe for more melinated folks to get in on UX design. It could literally be a life or death situation if we don’t.
Not sure what User Experience (UX) is? Here is a start. If you’re a Black person interested in getting more information on becoming a UX Designer- check this out.
