Further Reading — Chapter 4

The following resources are organized by accessibility. Start with the general-audience items and work toward more specialized material as your interest deepens.

General Audience

1. Cathy O'Neil, Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy (2016) A widely read introduction to how data-driven models can encode and amplify social biases. O'Neil — a mathematician and former Wall Street quant — examines case studies in policing, education, lending, and hiring. Highly accessible, requires no technical background. Directly relevant to this chapter's discussion of historical bias and feedback loops.

2. Virginia Eubanks, Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor (2018) Eubanks investigates how data-driven systems affect low-income Americans, examining automated eligibility systems for public benefits, predictive models in child protective services, and coordinated entry systems for homelessness. Excellent companion to the discussion of ghost data and who is missing from datasets.

3. Mary L. Gray and Siddharth Suri, Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass (2019) The definitive treatment of the hidden human labor that powers AI. Gray and Suri, both researchers at Microsoft Research, draw on extensive interviews with workers on platforms like Amazon Mechanical Turk and interviews with tech executives. Directly extends Case Study 2 in this chapter.

Intermediate

4. Joy Buolamwini and Timnit Gebru, "Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification" (2018) The research paper behind Case Study 1. Buolamwini and Gebru's methodology is straightforward and the writing is accessible even for non-technical readers. Published in the Proceedings of the Conference on Fairness, Accountability, and Transparency (FAccT). Available freely online.

5. Timnit Gebru, Jamie Morgenstern, Briana Vecchione, Jennifer Wortman Vaughan, Hanna Wallach, Hal Daume III, and Kate Crawford, "Datasheets for Datasets" (2021) The paper that introduced the concept of standardized dataset documentation. Proposes a structured set of questions that dataset creators should answer, covering motivation, composition, collection process, preprocessing, uses, distribution, and maintenance. Published in Communications of the ACM. A practical resource for anyone creating or evaluating datasets.

6. Kate Crawford and Trevor Paglen, "Excavating AI: The Politics of Images in Machine Learning Training Sets" (2019) An investigation into the ImageNet dataset's person categories, revealing how cultural biases and stereotypes get encoded into training data labels. Part research paper, part art installation, and entirely eye-opening. Available online at excavating.ai.

Specialized / Academic

7. Safiya Umoja Noble, Algorithms of Oppression: How Search Engines Reinforce Racism (2018) An examination of how search engine results reflect and reinforce racial and gender stereotypes. Noble's analysis of Google search results for terms related to Black women and girls demonstrates how data-driven systems can perpetuate harmful representations. More academic in tone but accessible to motivated general readers.

8. Ruha Benjamin, Race After Technology: Abolitionist Tools for the New Jim Code (2019) Benjamin coins the term "New Jim Code" to describe how AI and data-driven systems can encode racial hierarchies under a veneer of objectivity. The book examines healthcare, criminal justice, and everyday technologies. Theoretically rich and provocative, suitable for readers interested in the structural dimensions of data bias.

9. Abeba Birhane, "Algorithmic Injustice: A Relational Ethics Approach" (2021) An influential paper arguing that mainstream approaches to algorithmic fairness — which focus on individual data points and statistical metrics — fail to capture the relational and structural nature of injustice. Proposes an alternative framework grounded in relational ethics. Published in Patterns (Cell Press). More theoretical but highly relevant to the data ethics discussion in this chapter.

Multimedia and Interactive

10. Coded Bias (documentary, directed by Shalini Kantayya, 2020) A documentary following Joy Buolamwini's work on facial recognition bias, featuring interviews with researchers, activists, and policymakers. Available on streaming platforms. An excellent visual companion to Case Study 1.

11. Google's "Know Your Data" tool (knowyourdata.withgoogle.com) An interactive tool that lets you explore the composition and characteristics of common machine learning datasets. Useful for hands-on exploration of the representational gaps discussed in this chapter.

12. The Economist, "The data labellers: Last of the human-in-the-loop workers?" (2023) A journalistic investigation into the global data labeling industry, examining working conditions, pay, and the industry's future. Accessible and well-reported; a good supplement to the ghost work discussion.