An organization combining art, research, and advocacy to fight algorithmic harm — and proof that affected communities can produce the research that changes industry.
What It Is
Founded by Joy Buolamwini after her research revealed racial and gender bias in commercial facial recognition systems, the Algorithmic Justice League (AJL) works to:
- Research — Document algorithmic harms with rigorous methodology
- Advocate — Push for policy changes and corporate accountability
- Educate — Make algorithmic bias legible to affected communities
- Create — Use art and storytelling to reach broader audiences
The organization emerged from the Gender Shades study, which found facial recognition error rates up to 34% for darker-skinned women versus less than 1% for lighter-skinned men.
Why It Matters for Redistribution
Who benefits:
- Communities harmed by algorithmic systems — their experiences are documented and amplified
- Policymakers — rigorous research informs better regulation
- Companies — pressure leads to improvements (IBM exited facial recognition)
Who bears costs under the status quo:
- Black women and other marginalized groups — highest error rates, least voice in development
- Communities subject to surveillance — facial recognition deployed without consent
What changes: AJL demonstrates that affected communities can produce world-class research challenging corporate claims. The Gender Shades methodology is now industry standard. Multiple cities banned government facial recognition after AJL advocacy.
The model: Research without advocacy gets ignored. Advocacy without research gets dismissed. AJL combines both — rigorous academic work driving real policy change.
AJL's work sits at the intersection of several redistribution questions. The Gender Shades study didn't just find bias — it revealed how The Frontier Lab Oligopoly ships products that work well for some demographics and fail others, with no accountability mechanism. The pattern Virginia Eubanks documents in Automating Inequality (2018) — where automated systems impose the highest costs on the most vulnerable — is exactly what AJL makes visible and contestable.
Current Status
Active — AJL continues research, advocacy, and education initiatives.
Key impacts:
- IBM exited facial recognition market
- Microsoft called for regulation
- Multiple cities banned facial recognition
- NIST now requires demographic testing
- EU AI Act includes provisions influenced by this work
How to Get Involved
- Learn: Watch Coded Bias documentary
- Support: Donate to AJL
- Advocate: Push for algorithmic accountability in your community
- Research: Build on Gender Shades methodology
- Report: Document algorithmic harms through AJL's channels
Related
- Gender Shades
- AI Now Institute
- DAIR Institute
- Stochastic Parrots
- Datasheets for Datasets
- Democratic AI Governance
- The Digital Divide in AI
Sources
- Algorithmic Justice League. "About." ajl.org
- Buolamwini, Joy and Timnit Gebru. "Gender Shades." FAT, 2018. MIT
- Coded Bias. Directed by Shalini Kantayya. 2020. Documentary