No. 51 of 500 No. 15 in AI Ethics & Society

Automating Inequality

How High-Tech Tools Profile, Police, and Punish the Poor

By Virginia Eubanks

Print length
288 pages
Publisher
Picador
Reading level
Beginner
ISBN-13
978-1250215789
Published
August 6, 2019
Written for
Public-sector leaders, social workers, policy analysts and technologists building or buying government algorithms.

Synopsis

Political scientist Virginia Eubanks examines three automated systems that decide who receives help in the United States. In Indiana, a contract with IBM to modernize welfare eligibility replaced caseworkers with call centers, and the state denied about a million applications in three years, often for paperwork errors labeled "failure to cooperate." In Los Angeles, a coordinated entry system scores homeless people's vulnerability to ration scarce housing. In Allegheny County, Pennsylvania, a predictive model rates the risk that children reported to a hotline will face abuse or neglect. Eubanks spent time with caseworkers, administrators and the families on the receiving end of each system.

She places these tools in a long American tradition she calls the "digital poorhouse," arguing that data systems continue the work of county poorhouses and scientific charity by profiling, policing and moralizing about poor people. The Allegheny tool, which she treats as a best-case design, still troubles her because it relies on data that agencies collect mainly about families who use public services. Critics have asked whether her cases separate the harms of technology from those of underfunding and human bias. Readers gain detailed case studies and a set of questions about consent, data and accountability that apply directly to AI systems now used in benefits, policing and child welfare.

Scorecard

Editorial score

How it scored

Editorial score: 73.6 / 1005 weighted criteria

  • Influence30%

    24.0/30
  • Authority25%

    18.8/25
  • Practical value20%

    13.0/20
  • Currency15%

    9.8/15
  • Readability10%

    8.0/10
  • Editorial score

    73.6/100

How the rank was computed

Rank No. 51 of 500 comes from a composite of 69.1: observable signals 36.0 of 55 points plus editorial review 33.1 of 45 points.

All 11 inputs
InputPoints
Observable signals36.0/55
Reach (Amazon ratings volume)9.9/16.5
Consensus across best-of lists13.8/13.8
Awards and bestseller evidence5.0/8.3
Reader rating, sample-adjusted6.1/8.3
Amazon sales rank1.2/5.5
Recency0.0/2.8
Editorial review33.1/45
Author authority8.4/11.3
Influence10.8/13.5
Practical value5.9/9.0
Readability3.6/4.5
Currency4.4/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A humane, closely reported investigation of how public-sector algorithms ration aid, with a persuasive case that efficiency cannot replace adequate funding.

Strengths

  • Rich on-the-ground reporting with affected families
  • Three detailed, contrasting case studies
  • Strong historical framing of welfare surveillance

Keep in mind

It predates generative AI, and some reviewers found that it does not clearly separate technology's harms from those of underfunding and human bias.

Key takeaways

  1. Automated eligibility systems can turn minor paperwork mistakes into benefit denials, shifting the burden of system errors onto applicants.
  2. Predictive risk models trained on public-service records mostly see families who rely on public services, which can concentrate scrutiny on poorer households.
  3. Ranking people for scarce resources such as housing does not reduce scarcity, and better triage cannot replace adequate funding.
  4. Data systems aimed at the poor continue a long history of poorhouses and surveillance, so new tools should be judged against that legacy.

Notable facts

About the author

Virginia Eubanks

Virginia Eubanks is an associate professor of political science at the University at Albany, SUNY. She holds a PhD in science and technology studies from Rensselaer Polytechnic Institute, was a New America fellow in 2016-17 and wrote Digital Dead End (2011). She is a founding member of the Our Data Bodies Project and helped found the grassroots group Our Knowledge, Our Power.

Questions

What is Automating Inequality about?
Automating Inequality investigates how automated systems decide who receives public help in the United States. Virginia Eubanks reports on Indiana's automated welfare eligibility system, a Los Angeles registry that ranks homeless people for housing, and an Allegheny County model that scores child-welfare risk, arguing that such tools build a "digital poorhouse" that profiles and polices poor families.
Is Automating Inequality still relevant in the age of generative AI?
Yes. Governments are expanding the use of AI in benefits processing, fraud detection and child welfare, and the book's central questions still apply: what data feeds a model, who is watched most closely, how errors are appealed and whether automation substitutes for funding. Its case studies predate large language models, so pair it with newer work on current tools.
Who should read Automating Inequality?
It is written for general readers but is especially useful for public-sector managers, social workers, policy analysts and engineers who design or procure government algorithms. Anyone weighing a risk-scoring or eligibility system will find concrete examples of failure modes and the perspectives of the people affected, which vendor materials rarely include.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.