Further Reading: Bias and Fairness

Essential Reading

Books

Weapons of Math Destruction by Cathy O'Neil (2016) A former Wall Street quantitative analyst examines how mathematical models used in hiring, lending, policing, and education reinforce inequality. Written for a general audience, this book was one of the first to bring algorithmic bias into mainstream conversation. Particularly strong on feedback loops and the way opaque models affect the most vulnerable. Accessibility: No technical background needed.

Race After Technology by Ruha Benjamin (2019) A sociologist examines how technology — designed with the assumption of neutrality — can encode and amplify racial hierarchies. Benjamin introduces the concept of the "New Jim Code," drawing connections between historical racism and contemporary algorithmic systems. The book pushes readers beyond the "fix the bias" framing toward structural analysis. Accessibility: No technical background needed; assumes some familiarity with social justice concepts.

Algorithms of Oppression by Safiya Umoja Noble (2018) Examines how search engine algorithms — particularly Google — reflect and reinforce racism and sexism. Noble's research on how search results for terms like "Black girls" returned pornographic and degrading content illustrates how bias is embedded in the infrastructure of information. Accessibility: No technical background needed.

The Alignment Problem by Brian Christian (2020) A broader look at the challenge of ensuring AI systems do what we actually want them to do, with substantial sections on fairness and bias. Christian weaves together computer science, philosophy, and psychology to explain why specifying human values in formal terms is harder than it sounds. Accessibility: Accessible to general readers, though some sections engage with technical concepts.

Articles and Papers

"Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification" by Joy Buolamwini and Timnit Gebru (2018). Proceedings of the Conference on Fairness, Accountability, and Transparency. The landmark study discussed in Case Study 1. The full paper is freely available online and is remarkably readable for an academic paper. Worth reading for the methodology as much as the results. Accessibility: Moderate — some statistical concepts, but the main findings are clearly presented.

"Machine Bias" by Julia Angwin, Jeff Larson, Surya Mattu, and Lauren Kirchner (2016). ProPublica. The investigative journalism piece that broke the COMPAS story, discussed in Section 9.3. Available for free on ProPublica's website. The accompanying methodology document is a model of transparent data journalism. Accessibility: Written for a general audience.

"Dissecting Racial Bias in an Algorithm Used to Manage the Health of Populations" by Ziad Obermeyer, Brian Powers, Christine Vogeli, and Sendhil Mullainathan (2019). Science, 366(6464), 447–453. The study that found a widely used healthcare algorithm systematically underestimated the health needs of Black patients because it used healthcare costs as a proxy for health needs. A crucial example of measurement bias with concrete consequences. Accessibility: Moderate — the methods section is technical, but the introduction and discussion are accessible.

"Fairness and Machine Learning: Limitations and Opportunities" by Solon Barocas, Moritz Hardt, and Arvind Narayanan (2023). Available free online at fairmlbook.org. A comprehensive textbook on fairness in machine learning, available for free. Covers the mathematical foundations of fairness definitions, the impossibility theorem, and mitigation strategies. More technical than this chapter but written to be accessible to readers without advanced math. Accessibility: Moderate to advanced — assumes comfort with basic probability and statistics.

Deeper Exploration

On the Impossibility of Fairness

"Inherent Trade-Offs in the Fair Determination of Risk Scores" by Jon Kleinberg, Sendhil Mullainathan, and Manish Raghavan (2016). Proceedings of Innovations in Theoretical Computer Science. The formal proof that calibration and balance (a version of equalized odds) cannot be simultaneously achieved except in trivial cases. The paper is technically dense but the first few pages are accessible and clearly explain the intuition. Accessibility: Advanced — requires mathematical sophistication.

"Fair Prediction with Disparate Impact: A Study of Bias in Recidivism Prediction Instruments" by Alexandra Chouldechova (2017). Big Data, 5(2), 153–163. A companion impossibility result showing that calibration and equal false positive/negative rates are incompatible when base rates differ. Accessibility: Advanced.

On Structural Perspectives

"Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence" by Shakir Mohamed, Marie-Therese Png, and William Isaac (2020). Philosophy & Technology, 33(4), 659–684. Examines AI through a decolonial lens, arguing that patterns of colonial power — extraction, exploitation, and control — are being reproduced in AI systems. Pushes the bias conversation beyond "fix the model" toward questions about global power and justice. Accessibility: Moderate — academic but clearly written.

"Fairness Is Not Static: Deeper Understanding of Long Term Fairness via Simulation Studies" by Alexander D'Amour et al. (2020). Proceedings of the Conference on Fairness, Accountability, and Transparency. Demonstrates that one-time fairness interventions can have complex, sometimes counterproductive long-term effects. Important for understanding why ongoing monitoring matters. Accessibility: Moderate to advanced.

On Auditing and Accountability

"Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing" by Deborah Raji, Andrew Smart, Rebecca N. White, et al. (2020). Proceedings of the Conference on Fairness, Accountability, and Transparency. Proposes a practical framework for organizations to audit their own AI systems throughout the development lifecycle. Useful for thinking about how bias audits work in practice. Accessibility: Moderate.

Multimedia Resources

Videos and Talks

"The Coded Gaze: Bias in Artificial Intelligence" — Joy Buolamwini, TEDx talk (2016) A concise and compelling introduction to facial recognition bias, featuring Buolamwini's personal experience and research. Excellent for sharing with people who prefer video to reading. Duration: ~9 minutes.

"Coded Bias" — Documentary directed by Shalini Kantayya (2020) A feature-length documentary following Joy Buolamwini and other researchers and activists fighting algorithmic bias. Available on streaming platforms. Duration: ~85 minutes.

"21 Fairness Definitions and Their Politics" — Arvind Narayanan, tutorial at FAT (2018) A detailed walk-through of different mathematical definitions of fairness and the value judgments each embodies. More technical than this chapter but presented clearly by one of the field's leading researchers. Duration: ~60 minutes. Available on YouTube.*

Podcasts

"In Machines We Trust" — MIT Technology Review podcast series Multiple episodes exploring AI bias, fairness, and accountability across different domains. Various episodes, 20–40 minutes each.

Interactive Resources

"Fairness in Machine Learning" interactive course — Google's Machine Learning Crash Course A free, interactive module on fairness in ML, including hands-on exercises with fairness metrics. Duration: ~1–2 hours.

AI Fairness 360 — IBM Research An open-source toolkit for examining and mitigating bias in machine learning models. Includes tutorials and documentation. For readers interested in seeing how bias mitigation works in practice with code. Accessibility: Requires some Python knowledge.

How to Use These Resources

• Starting out? Begin with O'Neil's Weapons of Math Destruction and the ProPublica COMPAS article — both are accessible and compelling.
• Want the research? Read Buolamwini & Gebru (2018) and Obermeyer et al. (2019) — both are landmark studies that are more readable than typical academic papers.
• Ready for depth? The free online textbook by Barocas, Hardt, and Narayanan at fairmlbook.org is the most comprehensive resource on fairness in machine learning.
• Prefer video? Start with the Coded Bias documentary and Buolamwini's TEDx talk.