Several authoritative sources have raised the possibility that computer counting and monitoring of work in automated workplaces will transform offices into electronic sweatshops. This paper examines this idea from the vantage point of industrial sociology and managerial theory. Five theoretical models are developed, each of which generates hypotheses about the contexts in which work monitoring becomes important. A brief history of clerical work is given which shows the antecedents of surveillance and work-measurement in this sphere, and a case study of (...) control in an automated office is presented in order to illustrate certain practical limitations on the use of computer surveillance. These insights are combined into a synthetic model which describes those contexts in which surveillance would be likely to result in speed-ups, and those where it would not. (shrink)