Throughout history, Black people have been marginalized and excluded from many areas of society. This includes not only socioeconomic and political spheres but also technological advancements. From the digital divide that still exists in many parts of the world to the underrepresentation of Black voices in the tech industry, the exclusion is persistent. A stark manifestation of this bias is seen in the development and implementation of Artificial Intelligence (AI) systems.
AI is increasingly becoming a cornerstone of our world, influencing everything from the advertisements we see to the sentencing of criminals. But like any tool, it mirrors the biases of its creators. As Google's AI Ethics Co-Lead, Timnit Gebru, once stated, “AI is not only a tool that has been shaped by the society we live in but a tool that will shape the society we live in."1 With the underrepresentation of Black individuals in AI development, we risk fostering a new generation of technology that's tinted with unconscious racial bias.
In a 2019 study by the AI Now Institute at NYU, it was found that the AI industry has a staggering diversity crisis, with over 80% of AI professors being male and a majority being white2. This lack of diversity will lead to racial bias, which we've already seen in AI applications. From facial recognition software that misidentifies Black people at five times the rate of white people3 to biased predictive policing algorithms, the consequences of racial bias in AI can be severe and life-altering.
Historically, racial bias in technology has resulted in what Safiya Umoja Noble, in her book "Algorithms of Oppression," describes as "a new form of redlining"4. Similar to the mid-20th century practice of denying services to residents of racially defined neighborhoods, today's digital redlining can result in Black communities being overlooked for tech services and innovations.
Addressing this bias is crucial, and it begins with greater Black involvement in AI development. For too long, the voices of Black scholars, engineers, and entrepreneurs have been sidelined in tech conversations. Encouraging greater diversity within AI will usher in technology that truly serves everyone, not just the privileged few.
Involvement takes many forms. Studying computer science, data science, or AI, joining AI research groups or tech companies, and pushing for more equitable practices within these organizations are all ways Black individuals can get involved.
Diverse representation also helps generate new ideas, challenge existing bias, and push AI to its full potential. As the famous computer scientist Grace Hopper once said, "The most dangerous phrase in the language is, 'We've always done it this way.'"5 Bringing a wide range of perspectives to AI development ensures that we're not just doing things the way they've always been done - we're pushing boundaries, breaking down barriers, and creating an inclusive technological future.
We have to encourage initiatives aimed at promoting diversity within AI and tech, such as Black in AI, a group dedicated to increasing the presence of Black individuals in AI research and reducing bias in AI6. Similarly, educational institutions and companies need to make a concerted effort to attract and retain Black talent, addressing structural barriers and biases that may deter or discourage Black individuals from pursuing careers in AI.
Overcoming racial bias in AI is not a task for Black individuals alone. It's a societal issue that requires concerted effort and systemic change. But the involvement of Black individuals in AI development is a crucial step towards a more equitable technological future. As we look ahead, let's not just dream of a world where AI is free from racial bias. Let's take action to create it.
Sources:
- 1. "Google AI Tool Will No Longer Use Gendered Labels like 'Man' or 'Woman' in Photos of People". (2020, February 20). The Washington Post.
- 2. West, S. M., Whittaker, M., & Crawford, K. (2019). Discriminating Systems: Gender, Race, and Power in AI. AI Now Institute.
- 3. Buolamwini, J., & Gebru, T. (2018). Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification. Proceedings of Machine Learning Research, 81, 1-15.
- 4. Noble, S. U. (2018). Algorithms of Oppression: How Search Engines Reinforce Racism. NYU Press.
- 5. Hicks, M. (2018). Programmed Inequality: How Britain Discarded Women Technologists and Lost Its Edge in Computing. MIT Press.
- 6. Black in AI. https://blackinai.github.io/.