Imagine you have the superpower to make the rules for a new world, a world where gender, nationality, religion, race, caste, class, sexual orientation, or physical and cognitive disability pose no barrier. One simple way to do that would require looking at the rules of our present world that hinder equity for these vulnerable populations.
What kind of world would you design?
Now, let’s zoom into our current global reality. According to the World Economic Forum (WEF), the world generated about 44 zettabytes of data in 2020. World leaders and experts are already grappling with understanding the impact of Artificial Intelligence (AI) on human life and even free will. Whether we like it or not, we are entering a data-driven world. This breakneck pace begs one question – who is making the rules in this new world?
It seems the answer is, a relatively homogenous group of people: Only 15% of data scientists are female with a skew towards early career roles rather than managerial roles. Similarly, Black employees represent less than 5% of Google, Facebook and Microsoft’s workforce, significantly less when compared to their representation across the global population. These numbers are not likely to improve across regions or sectors.
When a workforce that is not diverse creates the rules, especially as we forge our path in this data-driven world, we move away from the idea of fair and equal – as seen through biases in hiring decisions, access to housing, policing practices, prison convictions, among other areas. For reference, revisit the infamous bias reported in Amazon’s AI-based recruitment which penalized applications with words such as “women”. These data-related biases can arise even before data is collected – while framing the problem – and can find place anywhere across from data collection, preparation, and beyond. These biases are hard to overcome because a non-diverse workforce often lacks the social contexts to properly identify and frame a problem. There is a business, moral, and social impact case to be made to reduce inequities. And the solution lies in diversity – of all kinds and at all levels.
A diverse workforce allows for different vantage points and closes the social context gaps that hinder unbiased decision-making. If we are to negotiate fairness and equity in a data-driven world, then it is critical to hire a diverse team.
Here are three points to keep in mind as you build your new, more diverse team:
- Measure and report: We must be better at measuring and reporting on the diversity of the global data-driven workforce. Presently, information about the lack of diversity in the data workforce is mostly US-centric. Limited information, if any, is available for other geographies, such as caste representation in India. Further, there is insufficient data in US reporting for people who do not identify as either a man or a woman, people from various sexual orientations, people of different religions, etc. Building a more diverse workforce will be difficult to advocate for without quantifying the problems and regularly measuring our global progress.
- Commit to dismantling the system: The lack of diversity is systemic and starts long before we enter the workforce. Unless we commit to change, it will continue. The data we measure and report must be leveraged to identify systemic barriers and develop solutions. For example, with data in hand, leaders and mentors can encourage others to aspire to a career in data, close the pay gaps, and build incentives for organizations to hire diverse talent and ensure representation at all levels – including in leadership. In addition to bolstering human resources, we must support policies that create a welcoming, inclusive infrastructure, such as breastfeeding rooms for new moms and health insurance for same-sex partners, to name a few.
- (Most importantly) Invite diverse voices and perspectives to build and create change: You will get the best results if you involve people with rich lived experiences and people who have been historically marginalized and excluded. Do not make the mistake of solving the problem for them. Solve the problem with them. Mere representation does not allow people from historically marginalized communities to exercise their agency. Question the intersection of power structures at play around you and if others feel confident to articulate their experiences. It is up to you to educate yourself and to be truly open to listening to diverse perspectives without dismissing or silencing them.
According to WEF, data scientist and data analyst jobs will have the highest growth in demand by 2025. There is no time to waste. Let’s seize this opportunity to harness our superpower and build an inclusive and equitable data-driven world.
This post has been written with research support by Young India Fellows, Ishi Shandilya and Khushi Baldota from Ashoka University.