The Algorithmic Justice League describes itself as “an organization that combines art and research to illuminate the social implications and harms of artificial intelligence.” Its stated mission is “to raise public awareness about the impacts of AI, equip advocates with resources to bolster campaigns, build the voice and choice of the most impacted communities, and galvanize researchers, policymakers, and industry practitioners to prevent AI harms.”
Origin
By AJL’s account, its founder Joy Buolamwini encountered the problem directly as a graduate student at MIT: facial analysis software failed to detect her face, while detecting her lighter-skinned peers and even a face drawn on her palm. She completed the project wearing a white mask so the software would register her. That episode prompted the research programme AJL was built around.
Research record
Buolamwini’s early work documented “large gender and skin type bias in commercially sold products from reputable companies including IBM and Microsoft.” She co-authored Gender Shades with Timnit Gebru, published in 2018 in the Proceedings of Machine Learning Research, and the follow-up Actionable Auditing with Deborah Raji, which examined Amazon’s system. AJL states that when Amazon sought to discredit the peer-reviewed research, more than 70 researchers defended the work, and that the National Institute of Standards and Technology subsequently published a study finding extensive racial, gender and age bias in facial recognition algorithms.
Method
The organization’s distinctive method is pairing peer-reviewed auditing with artistic work intended to make the findings legible outside the field — Buolamwini’s award-winning spoken word piece “AI, Ain’t I A Woman?” has been shown in exhibitions internationally.
Relevance to digital political discourse
AJL is a case of a research finding, rather than a campaign or a constituency, becoming the organizing basis for a policy argument: a measured accuracy disparity in commercial products turned into legislative and procurement debate over facial recognition. Its stated policy position is that accountability for AI systems must “go beyond self-regulation” and that those affected must have access to redress.
Sources
- 01.
Algorithmic Justice League. Mission, Team and Story. Source for AJL's quoted mission and self-description as "an organization that combines art and research to illuminate the social implications and harms of artificial intelligence"; for Joy Buolamwini's founding account, including the facial analysis software that failed to detect her face at MIT; for the Gender Shades paper co-authored with Timnit Gebru and the Actionable Auditing paper with Deborah Raji; for the "AI, Ain't I A Woman?" spoken word piece; and for the account of the dispute with Amazon and the subsequent NIST study.
- 02.
Buolamwini, J. & Gebru, T. (2018). Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification. Proceedings of Machine Learning Research, 81, 77-91. The peer-reviewed paper underlying AJL's facial recognition work, published in the Proceedings of the 1st Conference on Fairness, Accountability and Transparency.