Browse Books

Reviewer: Rosa Michaelson

The main intent of this collection of papers and brief commentaries is to highlight the need for consideration of human values in the design of computer technology. What does the phrase "human values" mean in this context__?__ In her introduction, the editor distinguishes between design for usability, the mainstay of human-computer interaction (HCI) research, and design that aims to mitigate detrimental effects on humans. Thus, "human values" covers all of the social stuff that computer science or software engineering is assumed to ignore-such as ethics, accountability, equality, and individual freedom-and is more commonly regarded as human-centered design, with all of the associated moral principals included. It is less clear for whom this book was intended: since contributors were mainly those with a background in HCI and computer-supported cooperative work (CSCW), those interested in these disciplines might be among the expected audience. Although Friedman explicitly wishes to broaden the approaches of computer science to design, the book seems to be more of a collection of teaching texts for a social context or ethics module in computing than a polemic for change. The book is arranged into three sections. The first, "Conceptualizing Human Values in Design," includes two very interesting papers. Suchman's paper "Do Categories Have Politics__?__" is a classic text. Here, she critiqued Winograd and Flores' use of speech act theory (after Austin and Searle) as a means for better software design in papers originally published in the 1980s. Suchman's critique concerns the power inherent in those who control the choice of categories used in the software design process, and the effects on the users of software produced under certain forms of categorization and assumptions about work processes, based on a modified form of speech act theory. Winograd observes, in response, that her argument is applicable to any systematization, as it is unlikely that there exists a method of producing software and hardware that does not involve categories and codification. Friedman takes up the idea of power and category in the introduction, worrying whether an editor can be democratic in choosing items and order, and still maintain coherence. This seems to me to be a metalevel problem of the domain, which is addressed more succinctly by other disciplines such as philosophy. Categorizing is not intrinsically nondemocratic, nor is a democratic process a necessary condition for a more humane outcome, though such a process seems more likely to produce computer systems that take account of the range of needs of all of the users. However, the way speech act theory was used (by academics, if not in "real-life" software engineering) to justify managerial views of work had an important impact in the late 1990s [1]. The first section also includes an interesting essay on the need for accountability in software design, and another paper discusses disability issues in design. I found this of great help in teasing out how exclusion from certain forms of technology may occur. However, noninclusion happens to several other groups of society, such as the elderly and women, whose particular concerns are not covered in the book, despite the ideas of cyberfeminism, which were debated at the time. The second section, "Computers as Persons__?__ - Implications for Design," concerns anthropomorphism and computers. Several chapters discuss the need of some people to treat inanimate objects as if they were human; some authors go so far as to regard computers as "social actors." Luckily, sense prevails in the conclusions of several of the items here; the worry is that moral responsibility and decision making will be ascribed to computing technologies by users. The early artificial intelligence debate about Eliza covers much of the same ground-research into avatars, virtualization, or the claims of those who think they have created emotional software would be more appropriate as a focus for teaching social values in 2005. The third section deals with the practice of value-sensitive design. The four papers in this section are case studies: a hospital database for portable equipment maintenance, a redesign project with economic restraints that failed to get rid of a superfluous microphone on-off switch, the implications of proposed car-tracking systems, and accessibility issues for a telephone company. The paper on privacy and surveillance issues associated with proposals for car-tracking systems has been overtaken by improvements in satellite technologies and the need for congestion control in cities. Privacy and surveillance issues are more usefully illustrated by the literature on the widespread use of outsourced call center technologies, or systems-integration plans for identity cards and passports. Access for the disabled to new technologies is still a hot topic-how many mobile phones have interfaces that the blind can use__?__ If this book were a collection of exemplary research into human-value design, in this section one might expect to find examples of the use of methods developed by socially aware designers, such as Mumford [2], Kling [3,4], and Checkland [5], but these are conspicuous by their absence. Ten of the 17 labeled chapters were published in journals and conference proceedings between 1992 and 1996, so timelines and relevance of the contents are of interest to the reader. Certain aspects of this book are still important, namely the way that the writings illustrate ambiguities inherent in this area of research. These include issues such as the problem of separating the technological constraints from the social context in design, the reification of terms such as "systems" and "technology" (which leads to a dearth of practical and technically based focus) [6], the unpredictability of use, and the impossibility of designers second-guessing emergent features of complex systems. Add in the sweeping generalities systems researchers often produce from unreproducable case studies (often of simple forms of particular technologies in localized contexts), and there are endless possibilities for academic discourse concerning computer systems design. As a resource for teaching about the social aspects of computing, this book contains many useful items. As a historic reference for current research, it is also of use, but there are gaps in its coverage of social and moral problems, including issues of gender, globalization, and the effect of the Internet. I found the first and third sections of this book to be of most lasting interest, and felt that the second has not aged well. As a European, I found the overall lack of explicit political references telling. The word "capitalism" was not used anywhere, despite the stress on the social contexts of design in most of the writing. Cursory references to open source methods or participatory design were not enough for completeness concerning human values in design. Online Computing Reviews Service

Reviewer: Michael Lee Gordon

As editor of this collection, Batya Friedman has published a significant work that brings together articles on computers and human values. As computer technology becomes ever more ingrained in the fabric of society and humanity, there is a clear need to examine the impact of this trend on human values and the human condition. Articles by contributing authors address such topics as accountability in a computerized society; engaging public participation in information technology design; social human-computer interaction; interface agents; human agency and responsible computing; and social choice about privacy. These are provocative and thought-provoking topics. This book is interesting and important in that the articles address topics that can help people keep the humanity in the interaction with computers and help ensure that computers are instruments used by people, without discounting the quality of goods and services that result from the use of the computers. Leaders in industry and government who shape our societal future are encouraged to read this book. The chapter on computer bugs and accountability is alone worth the trip (or click) to the bookstore.