algorithmic-bias

Chapter 4, "The Allegheny Algorithm," presents an extensive compilation of interviews and published literature focused on the development and application of predictive risk modeling in child welfare within Allegheny County. It highlights the Allegheny Family Screening Tool, a data-driven algorithm designed to identify children at risk of maltreatment, alongside critical perspectives on the ethical, social, and systemic implications of using such technology in public welfare systems. The chapter situates this case study within broader discussions of data analytics, social policy, and the challenges of addressing inequality through automated decision-making.

This chapter presents a collection of endorsements highlighting the critical examination of how automated systems and digital technologies perpetuate and exacerbate social and economic inequalities, particularly affecting marginalized communities. The endorsements emphasize the book's exploration of the surveillance state, the punitive nature of digital welfare systems, and the urgent need to understand and resist these technological forces to promote justice and equity.

Chapter 5 of Eli Pariser's "To Save Everything, Click Here" critically examines the risks and consequences of algorithmic gatekeeping in digital platforms, highlighting how algorithms shape information visibility, influence public discourse, and can perpetuate biases or censorship. The chapter draws on various case studies and scholarly references to illustrate how algorithmic decisions impact cultural production, media representation, and democratic engagement, emphasizing the opaque and often unaccountable nature of these automated systems.

Chapter 3 primarily addresses foundational and contemporary challenges in artificial intelligence, referencing seminal works like Alan Turing's 1950 paper on machine intelligence and highlighting practical issues such as algorithmic bias and adversarial attacks in AI systems. The chapter also touches on technical methods like Monte Carlo tree search and notes regional differences in AI development and regulation.

The 'Preamble' chapter of Weapons of Math Destruction primarily consists of critical acclaim and publication details rather than substantive content from the book itself. It compiles numerous endorsements and reviews that highlight the book's examination of the pervasive and often harmful influence of algorithms and big data on society, emphasizing its role in increasing inequality and threatening democratic processes. The chapter sets the stage for the book's urgent critique by showcasing the broad recognition of its importance and relevance.

Chapter 5 of 'Weapons of Math Destruction' examines the use of predictive policing technologies such as PredPol and facial recognition software in various U.S. cities, highlighting their roots in historical policing strategies like zero-tolerance and stop-and-frisk. The chapter critiques these algorithms for perpetuating racial biases and inefficiencies, noting the disproportionate targeting of minority communities and the problematic incentives of private prisons. It also references studies and legal challenges that question the effectiveness and fairness of these data-driven policing methods.

Chapter 6 of "Weapons of Math Destruction" examines the use and fairness of workplace personality tests in hiring practices, highlighting their widespread adoption and potential biases. It discusses legal and ethical concerns, such as the classification of these tests as medical exams under the ADA, and explores how algorithms and data-driven tools influence recruitment, often perpetuating discrimination and inefficiencies. The chapter also references case studies, expert interviews, and research illustrating the challenges and consequences of relying on such opaque and unregulated assessment methods.

Chapter 8 of 'Weapons of Math Destruction' provides an extensive bibliography and references related to credit scoring, data brokerage, and the financial industry's use of algorithms. It highlights the pervasive influence of credit scores like FICO on consumer access to credit and employment, the racial wealth gap exacerbated by these systems, and the rise and challenges of alternative lending platforms such as peer-to-peer lending. The chapter also touches on the risks of algorithmic bias, data inaccuracies, and the opaque nature of data-driven decision-making in finance.