Ain’t I a Woman?

Ain’t I a woman? Four simple words combine to form a powerful question that provokes our sensibility. First attributed to Soujourner Truth in 1851, and then used in 1981 by feminist author and cultural critic bell hooks, these words were employed to challenge the way women, and particularly Black women, were discounted and made to feel invisible. Lest we think these problems are historic, I’d like to share the experience of MIT researcher, poet and computer scientist Dr. Joy Buolamwini who was compelled to pose the same question in her spoken word performance, “Ain’t I a Woman?” in 2015. 

Joy has a PhD from MIT. She is a Rhodes Scholar, Fulbright Scholar and Founder of the Algorithmic Justice League who found that no amount of high academic achievement protected her from the discriminatory power and frightening social implications of artificial intelligence. As a graduate student at MIT, Joy discovered that facial recognition software could not see her unless she donned a white mask. Her subsequent research revealed similar gender bias when AI software developed by leading companies determined that Oprah, Michelle Obama and Serena Williams were male. In other words, Black women were not women. 

In an article written for Time, Joy explains why such obvious errors occurred. She writes, “Less than 2% of employees in technical roles at Facebook and Google are Black. At eight large tech companies evaluated by Bloomberg, only around a fifth of the technical workforce at each are women.” Joy found one government dataset used to train computers in facial recognition “contained 75% men and 80% lighter-skinned individuals and less than 5% women of colour [sic].” In other words, the ways in which AI was being trained resulted in systems with inherent bias. 

Some may argue that such errors are simply to be expected as we work out the bugs inherent in a new technology. I argue that these types of errors are foreseeable and are not merely problems of code. They are problems of people who are fallible and problems of systems that limit which faces are at the table (or in the dataset). 

I worry that we sometimes intellectualize such problems and view them as intriguing case studies to discuss with our colleagues or teach in our classrooms and business schools. We readily separate the problem from the impact, thereby allowing the discussion to ignore the fact that real people are being hurt. Of course there is important learning to be had from the failures of any system but our ability and willingness often to depersonalize such moments and to locate them in the realm of academia or business strategy fail to acknowledge the victims. I cannot fathom what it feels like to be told I am not who I know I am by an algorithm. I cannot imagine what it feels like to know that I am misclassified because for generations past I have not been seen. 

We think of data as neutral and perhaps one could argue that it is when decontextualized. But data employed in the service of any task, particularly a coded task, takes on the bias of the coder. Joy’s organisation, the Algorithmic Justice League, as part of its mission, declares: “We want the world to remember that who codes matters, how we code matters, and that we can code a better future.”  

I hope that in our work as a school we never allow the excitement of discovery or the promise of new technologies to go unchallenged. We have a unique opportunity to help our exceptional students understand that everything we do has an impact, not only in the future but today. Our work, our words, our actions, our art, our code, our research – all of these things matter. Let’s ensure that at UTS, they matter for good. 

NOTE:  Last week we recognized the UN Day for the Elimination of Racial Discrimination. The theme of this year focuses on people of African descent, who, to quote UN Secretary António Guterres, “face a unique history of systemic and institutionalized racism, and profound challenges today.” These words made me think back to Joy Buolamwini’s research and the ongoing work of her organisation.  I would like to acknowledge Guterres’ important words as the impetus for this week’s message and as a challenge to our entire UTS community to consider how we can work together to create an inclusive and caring community, free from harassment and discrimination.

Read more of Dr. Foster's Blog

#fosteringbrilliance