But first, what are the differences between equity, diversity and inclusion?
Equity: "A principle that is based on a sense of what is just and what is unjust, above and beyond legal norms.”
Diversity: "The quality of a heterogeneous group of people who, in a given environment, differ from one another by characteristics that are generally social, cultural, physical or psychological.”
Inclusion: "Inclusion is about community. It is about creating a culture that promotes equity and celebrates, respects, accepts and values difference. The goal of diversity and inclusion is to capture the uniqueness of the individual, to create an environment that values and respects people for their talents, skills, and abilities they bring to the community. "
Inclusion is seen more as a state of mind. It is much harder to achieve than equity and diversity because everyone must have a voice and be heard.
Biases in artificial intelligence systems
When discriminatory bias is introduced in data segments from a population in artificial intelligence systems there is a strong possibility of a negative impact on society as it can easily fail to represent a large number of people at once. Certain times bias can result in missing a particular segment of the population. For example, when a system is built that fails to include historical data on people of colour. As a result this system is bound to make more errors for that micro segment of the population as it wouldn’t take it into account.
Another possible source of bias is part of the representation. For example, in designing recruitment software, what really matters are the skills of the candidates. Demographic information should not be important. It should not be necessary to know the birthplace, gender, ethnicity and religion of candidates. Organizations can claim that they will not use this information, but it is much more complex than that, as many other attributes have been indirectly coded into the demographic segment. This poses a big problem: it's like a bug in software.
All human-led processes are biased. For example, a woman applies for a mortgage and has to talk to a bank officer. It is possible that the bank officer is racist or sexist and she may not get the loan simply because the bank officer is biased. Why can this have serious consequences for the AI system? It is because a bank officer can deal with a maximum of one or two dozen people a day. Whereas thousands of people a day can be affected by a biased AI system. The impact of such a system is therefore much greater than if there were a racist or sexist individual in a financial institution. This is why better governance and quality assurance for such systems are essential.
The Impact of Lack of Diversity in Teams
Society benefits from the coming together of divergent viewpoints, as do companies, which perform better when their teams are diverse. According to Julia Elvidge, a retired engineer and former president of Chipworks, without diversity in teams:
"You don't get better decision-making, problem solving and good discussions. You can also look at diversity in terms of Myers-Briggs personality types. There are rational people and others who are emotional. There are extroverts and introverts, so look for this type of diversity in your team as well. The more diversity you have, the more ideas you have, and the more you can mix them together and decide which one is the right one. You want to avoid a group thinking approach where everyone has the same idea."
It can be diversity on boards, in management, or even in entry-level positions.
In addition, greater diversity in technology sector teams promotes innovation, economic performance and a better product offering. According to Eli Fathi, CEO of MindBridge.Ai and a two-time University of Ottawa alumnus in Electrical Engineering :
"When you consider the issue of technology and bias, currently less than 11% of women work in the Artificial Intelligence (AI) sector. In AI, you learn from the data. When you don't have women on the team, the number of biases increases dramatically in the results. For this reason, your product is not as good as it should be. Now the question is how many women are available in the labour market. If all companies had diversity at the management level, North America's GDP would rise to $1.5 trillion. The numbers are staggering. "As a result, if more women were to become engineers, it would provide a solution to the labour shortage."
Therefore, if more women became engineers, it would provide a solution to address the labor shortage.
