Reflecting on 'Data for Black Lives II'
Earlier this month, AISP’s Director of Training Amy Hawn Nelson and MSW intern Emily Berkowitz attended Data for Black Lives II (D4BL)—a two-day conference on the MIT campus in Cambridge, Massachusetts. Over 200 data scientists, programmers, activists, and elected officials came together to share and learn about “the use of data science to create concrete and measurable change in the lives of black people.” Yeshimabeit Milner, D4BL cofounder and executive director, kicked-off the convening by expressing that D4BL “is about using the datafication of our society as an opportunity to make bold demands for racial justice” and bring in “an era where data will be recognized as a tool for profound social change.”
The conference centered the voices of black communities, especially black women, in part to negate the belief that white dominance in these fields is caused by a lack of black expertise. “Our conference is living proof that there is no shortage of black technologists and data experts to speak about these issues,” Milner said. “If you cannot find black panelists for your conference, that means you are not looking.” The speakers and panelists were 85% women and gender non-conforming people, 82% black, and represented multiple professions—from organizers against police brutality to federal agency staff.
Keynote speaker Meredith Broussard began the conference by challenging the idea that simply having access to data will effortlessly create an equitable society, emphasizing that both individual prejudice and oppressive systems are reflected in data points because data are created by humans. Systems that have been fueled by racist policies, such as redlining and state-sanctioned racial exclusion by credit rating agencies in the United States, now live within pervasive machines and technologies. The panelists from We are the Leaders We Have Been Looking For: Organizing for Algorithmic Accountability (pictured above) all agreed: “Algorithms are not new; they are manifestations of old systems that devalue and dehumanize our people.”
To bring widespread attention to the damage caused when technological knowledge is privileged over other forms of meaning and intellect, Broussard coined the term technochauvinism. At its most explicit, technochauvinism is the valuation of technology over people, but it also takes a more subtle form in many of the social sciences and tech spaces when researchers and governments trust isolated quantitative data over qualitative data, context, and narrative. Tamika Lewis of Our Data Bodies, a D4BL panelist, encouraged audience members to ask, “What are the stories that folks want to share with us that go with statistics we see in reports?” and connect with the humanity behind the data points—a good tool for data experts to expand their work beyond numbers.
D4BL II demonstrated that data can be used for good too. Organizing for racial justice using data collection has long been a part of the civil rights movement. Ida B. Wells is regarded as a data pioneer; in the 1890s she developed statistics demonstrating the pervasive number of lynchings in the United States and reshaped the debate about racial justice through her journalism and reporting on the subject. As technology continues to evolve and both over-surveillance (i.e., disproportionate policing in majority black communities compared to majority white communities) and erasure of black communities (i.e., the systematic exclusion of black people in math, climate advocacy, etc.) proliferate, the ways data and technology are harnessed by activists and organizers also have to change. For example, Joy Buolamwini, founder of the Algorithmic Justice League, shared her work as a poet of code that reveals gender and skin-type bias in facial analysis technology. Conference workshops also empowered attendees to identify injustices and generate change by providing training on everything from the Freedom of Information Act to analytic tools like R and ggplot to create data visualizations.
The conference also touched on the role of data in politics and power. In the Local Organizing and Movement Building; Who Counts? workshop, Dejuana Thompson and Deanna Reed emphasized the importance of organizing African American voters in the South through local, community-led spaces, and accurate representation in voter mobilization efforts. Typical voter forecasting model inputs, Thompson explained, use voter file data, contact data, and commercial data, but fail to include community organizations, affiliations, and strongholds like churches, where African American voters are more likely to be represented. “Data tells a story,” Thompson’s slide read, “Root data in the culture.”
Thompson founded “Woke Vote”--a program that helped propel a Democrat from Alabama, Doug Jones, to the U.S. Senate for the first time in 25 years; 98% of black women voters—who are undervalued by the aforementioned measurement practices—cast their ballot for Jones, truly securing a dramatic win for Democrats. Thompson further stressed that it is not enough to be present within a community during election season; to be “about community” requires sustained investment and attention from politicians after they enter office.
Effective community involvement—particularly of black communities who have long been denied access to the table—requires accessible, shared language and dialogue about data and data-driven systems. Towards that end, Actionable Intelligence for Social Policy and the Future of Privacy Forum—which houses the ADRF Network—recently published Nothing to Hide: Tools for Talking (and Listening) About Data Privacy for Integrated Data Systems—a toolkit to help data integration efforts engage with their communities about data sharing efforts and build trusting partnerships. Read the report here.
At the conference, Milner also shared an innovative work-in-progress: a “nutrition label" that helps anyone assess if an algorithm is biased or not, reaching audiences outside the tech space (featured on the left). Ken Steif, Director of the Urban Spatial Analytics Program at the University of Pennsylvania School of Urban Design, and colleagues are working on a complimentary, soon-to-be-published project to address and easily translate algorithmic impact and fairness. It includes, among other things, a “scorecard” that standardizes how to assess these aspects across government algorithms.
The interdisciplinary and inclusive approach of D4BL II was accentuated by the centering of black women’s voices, knowledge, and experience. It was an incredible opportunity for AISP and the ADRF Network to expand its intersectional approach to data sharing. Thank you to the organizers, volunteer, panelists, and speakers for sharing their wealth of knowledge and fostering a welcoming, energized community. We are looking forward to Data for Black Lives III, and hope to see you there.
Search #Data4BlackLives on Twitter for more recaps and media from the conference.