Bias exists when it’s unchecked. That’s why homogenous tech development teams can sometimes be a problem. But even when you have diverse teams, if your cognitive load is high you can still make mistakes. The qualities of communication skills, creativity and innovation and attention to detail are essential in the technology industry, but they all require conscious consideration of diversity to improve accuracy.
We are currently working with several tech companies on how to incorporate inclusion into the product design and coding process. During our work with these teams, we spoke about this conundrum: the need to balance speed and accuracy. When speed is paramount, and we don’t have diverse perspectives ensuring calibration of our decision-making, problems can result.
In our haste, it’s all too easy to fail to notice and challenge unconscious bias, and even racist terminology/jargon in order to get the job done. But what do tech jargon terms such as “black hat”, “white hat”, “slave and master servers” reveal about bias in the industry?
In conversations with our software developers and tech product designer clients, it became clear to us that to be a truly excellent coder, aside from having the technical knowledge and skills necessary for your work, there are three things you need:
- Communication skills. Nobody is working alone anymore, even in coding. Everyone works in teams, so how you communicate matters.
- Creativity and innovation. Unless you’re at the lowest levels, your job is not just rote. You’re effectively solving problems all day, and sometimes you need to be creative to come up with a solution (and as a team you need to be creative together).
- Patience and logic. Code is just a set of logical commands at the end of the day. You need to be able to think through the very clear and strict logical steps of your solution in order for it to work the way you want.
All these skills can be threatened by high cognitive load. Missing small details can cause big problems: An errant semicolon can break the code and make it not run; or a missing tab can completely change the functionality of a program or algorithm.
Consider the example from one of our clients about how a classifying algorithm error led to their family internet security solution misclassifying LGBT resources for youth as pornography. Or when a broadcaster wrongly labeled the image of a Black comedian, and in a news clip the wrong person was played. These things happened because of bias.
Making the unconscious conscious
As we said, the qualities of communication skills, creativity and innovation and attention to detail are essential, but they all require conscious consideration of diversity to improve accuracy.
If you have these qualities and the technological skill, but your cognitive load is too high, then it’ll be harder as a coder to reach your full potential and bring as much to your work as you could otherwise.
If you’re Black, you’re already a rare talent in tech workplaces, particularly among programmers. You may be already in a slightly heightened state of awareness of your own race and the implicit biases others have. Add to this the current global situation and that awareness is even more front-of-mind. Your cognitive load is high.
Changing the language we use
Tech language was developed back in the early days of modern computing during a time when globally racism was much more explicit and often went unchallenged. But there is no reason we can’t change that language. It’s not embedded in the code itself; it’s just how we talk about these concepts.
I recently heard of an example where a team of coders working on a solution had to go through the “blacklist and whitelist” of terms/commands for a specific product. The “blacklist” was terms/commands they couldn’t or shouldn’t use while the “whitelist” was stuff that’s OK. Because of the Black Lives Matter movement and what’s in the news, they noticed these terms in a new light for the first time and changed the language they were using to avoid using those racialized terms. It’s easy to just use different words, so why not? It’s an easy low-cost, low-tech solution to change language and improve output.
More diversity, more inclusion
Recently, Microsoft removed terms like these from their documentation. Cloudflare is debiasing some of the terms used in their coding. There are no reasons why such simple conscious actions can’t be undertaken for the benefit of us all.
The benefits of diversity are widely stated. But they’re actually only available to companies when they include people. Changing the language used might seem minor, but it can make a big difference. It’s practical, concrete and embedded in the work of what the organization actually does.
In addition, many tech firms are now setting targets for greater diversity. Microsoft, for example, wants to double its number of Black executives by 2025. Wells Fargo has linked increased diversity in teams to corporate pay. But these efforts need to be consolidated by inclusion efforts and becoming more conscious of bias in the jargon that forms our everyday work interactions is a great place to start.
Stephen Frost is the founder and Raafi-Karim Alidina is an associate of global diversity and inclusion at consultancy Frost Included. They are co-authors of a new book Building An Inclusive Organisation, published by Kogan Page.