Reading & Research
Curated books, papers, films, podcasts, and organisations for understanding AI through a lens of race, justice, and accountability.
Start Here
The books and papers that built the intellectual foundation for this field.
How search engines reinforce racism. The foundational text for understanding how seemingly neutral technologies encode racial and gender bias. Essential reading.
Publisher page →A memoir and investigation into the harms and biases of AI — part personal story, part rigorous research. Written by the founder of the Algorithmic Justice League.
Publisher page →Introduces the concept of the "New Jim Code" — how seemingly neutral technologies can perpetuate racial discrimination under the guise of innovation and objectivity.
Author page →A rigorous and accessible framework for thinking about power in relation to data. Examines how the work of data science, including AI, can be approached from a feminist perspective. Full text available free online.
datafeminism.io →How technology intersects with democracy in Kenya and East Africa. A critical corrective to Western-centric narratives about technology and political participation.
Publisher page →Maps the material, political, and social conditions that make AI possible — mines, data centres, human labour, state power. Challenges the idea that AI is immaterial or neutral.
Publisher page →The landmark study that exposed racial and gender disparities in commercial facial analysis AI — up to 34.7% error rate for darker-skinned women vs 0.8% for lighter-skinned men. Changed the industry.
Read paper →The paper whose suppression by Google led to Timnit Gebru's firing. Examines the risks and harms of large language models. One of the most important AI safety papers written by women of colour.
Read paper →Proposes a standard documentation framework for datasets used to train AI systems. A practical tool for accountability that has influenced the whole field.
Read paper →The framework that underpins DataBias. Argues that when policy centres those at the sharpest intersections of oppression, it works better for everyone. Originally an economic policy framework, it applies directly to AI design.
Read report →Watch
For when you want to watch rather than read — and to share with people who aren't ready for an academic paper.
A landmark investigation into how OpenAI's pursuit of artificial general intelligence has reshaped power, labour, and society globally — from Silicon Valley to the Global South. Hao spent years reporting from inside the AI industry. Essential for understanding not just what AI does, but who it serves and who pays the price.
Publisher page →Follows Joy Buolamwini's research into facial recognition bias and the broader fight for AI accountability. The most accessible introduction to algorithmic justice for general audiences.
Film website →Buolamwini's landmark TED talk on algorithmic bias — combining spoken word poetry with research. Over 1.7 million views. Her follow-up 2022 TED AI talk on protecting human rights in an age of AI is equally important.
Watch on TED →Noble on how internet search can encode and amplify biases — and why that matters for democratic society. Clear, accessible, and short enough to share.
Watch on TED →A podcast/video series that applies rigorous critique to AI hype and misinformation. Funny, rigorous, and a useful antidote to the "AI will solve everything" discourse.
DAIR website →Listen & Subscribe
For staying current without reading every paper.
A newsletter that examines feminist approaches to data in a digital age. One of the inspirations for DataBias. Thoughtful, well-researched, and non-US in its perspective.
Read →The issue of Logic Magazine created in the aftermath of Timnit Gebru's firing from Google. A landmark document for understanding the politics of AI ethics and the cost of speaking up.
Read →Conversations from the Distributed AI Research Institute — community-rooted AI research, free from Big Tech influence. Subscribe via DAIR's newsletter to get episodes.
DAIR website →Talks and discussions from the Black in AI community — one of the most important networks for Black AI researchers globally. Workshop proceedings and talks are publicly available.
blackinai.github.io →Connect
Where to find the community, the research, and the advocacy.
Independent, community-rooted AI research free from Big Tech influence. Focus on marginalised communities, Africa, and African immigrants in the US.
dair-institute.org →Uses art, research, and advocacy to highlight the social implications and harms of AI. One of the most influential organisations in AI accountability.
ajl.org →Policy-focused research on the social implications of AI. Publishes annual reports and policy briefs. Based at NYU.
ainowinstitute.org →A network of 1,500+ Black AI researchers. Runs annual workshops at NeurIPS and offers networking and collaborative opportunities globally.
blackinai.github.io →A network for those working at the intersection of data science and feminist practice. Conferences, resources, and community.
datafeminismnetwork.org →UK-based independent research and deliberative body with a mission to ensure data and AI work for people and society. Important for UK policy context.
adalovelaceinstitute.org →Works to advance civil rights protections in the AI era. Focuses on algorithmic discrimination and the intersection of race, law, and technology.
aiforthepeople.org →Mechanism Design for Social Good — and its associated ACM conference on Equity and Access in Algorithms, Mechanisms, and Optimization. Rigorous, community-oriented, global.
eaamo.org →In Practice
Practical resources for applying these ideas in research and practice.
A documentation framework for machine learning models — transparently reporting a model's intended use, training data, performance across demographic groups, and known limitations.
modelcards.withgoogle.com →A database of AI system failures, harms, and near-misses. Essential for understanding what actually goes wrong with AI in deployment — not just in theory.
incidentdatabase.ai →Produces rigorous, accessible research on the social, cultural, and ethical dimensions of data-centric technologies. Reports, case studies, and policy briefs available free.
datasociety.net →A free course on data reasoning and how to critically evaluate AI claims. Useful for building the skills to interrogate AI research and media coverage.
callingbullshit.org →Help Grow This List
This list is maintained by the DataBias community. If you know of a book, paper, tool, or organisation that should be here — especially from outside the US and UK — please let us know.
Suggest a Resource →"The site will host works from academics and non-academics about topics related to data, algorithms, and AI — allowing an evaluation of current data policies." — DataBias original proposal, KCL, 2022