Technically Wrong

Sara Wachter-Boettcher’s Technically Wrong examines how routine digital products quietly define who counts as a “normal” user. The examples named by the publisher are deliberately ordinary: sign-up forms that fail people who are not straight, social platforms that send upbeat “year in review” messages about dead relatives, chatbots that harass women, and algorithms that put more Black people behind bars. These failures differ in scale, but the book links them through a shared design question: whose life did the product team imagine, and what happens to people the interface or model did not anticipate?

That question shifts analysis away from the comforting idea of a lone bug. A form that only offers a narrow set of gendered or family categories makes an institutional assumption look like a user’s personal complication. An automated memory feature can treat grief as just another opportunity for engagement because the system has no meaningful account of context. A predictive tool can translate historical inequality into a technical score and then make that score appear objective. In each case, a design choice distributes friction, visibility, and risk unevenly. The interface is not the whole cause; it is where deeper decisions about data, testing, product goals, and responsibility become tangible to the person using it.

Wachter-Boettcher connects these product failures to the values and processes of the technology industry. Her analysis challenges the belief that good intentions or a few more diverse hires will, by themselves, correct a system built around narrow assumptions. Representation matters, but it has to come with authority: who chooses the research questions, which harms count as launch blockers, who can stop a release, and whether affected users can shape the remedy. “User-centered” language is empty if organizations only listen to people who already fit the product’s imagined default. Ethical design is not a final audit added after the interface is finished; it is a continuing responsibility shared across research, engineering, content, leadership, and business decisions.

The book is especially effective at making structural critique legible through familiar interactions. Readers can recognize the small exclusions—a field that will not accept a name, a service that assumes one kind of family—and then trace them to wider patterns of organizational power. That movement from everyday friction to institutional accountability makes the book useful beyond design teams. It gives readers questions for evaluating products: what categories have been built in, whose experience is missing from the training or testing data, who bears the cost of an error, and what recourse exists when a company’s system causes harm? An algorithm’s output is not made neutral simply because it is automated.

The different stakes of the examples should not be flattened. A rejecting form can impose repeated everyday labor and signal that a person does not belong; a predictive score used in criminal justice can affect liberty on a much larger scale. The comparison is not that these harms are equivalent, but that both expose a chain of decisions that organizations often conceal behind the word “technical.” Following that chain means asking what the system optimizes, what it records as success, which people test it, and who has power to correct it. This turns inclusive design from a checklist of fields into an accountability question.

The record here represents the 2018 paperback (ISBN 9780393356045); the book first appeared in 2017. Its examples belong to that period, before generative AI became a mass-market product. The book is therefore a diagnostic foundation, not an audit of contemporary large language models or every current platform. Its enduring feminist contribution is a way of following responsibility from the micro-interaction back to the institution that designed it. For a broader account of computing’s entanglement with colonialism, race, gender, environment, and labor, pair it with Your Computer Is on Fire; for a more practice-based question about whether feminism can inhabit software architecture and interfaces, read Feminist in a Software Lab.

Sources: W. W. Norton’s paperback record (ISBN, edition date, and examples in the publisher description); Sara Wachter-Boettcher’s book page (author’s framing of bias and blind spots in digital products).