Two alternative work designs are identified for operators of stand-alone advanced manufacturing technology (AMT). In the case of specialist control, operators are limited to running and monitoring the technology, with operating problems handled by specialists, such as engineers. In the case of operator control, operators are given much broader responsibilities and deal directly with the majority of operating problems encountered. The hypothesis that operator control would promote better performance and psychological well-being than would specialist control (which is more prevalent) was tested in a longitudinal field study involving work redesign for operators of computer-controlled assembly machines. Change from specialist to operator control reduced downtime, especially for high-variance systems, and was associated with greater intrinsic job satisfaction and less perceived work pressure. The implications of these findings for both small and large-scale applications of AMT are discussed.One of the most important issues currently facing manufacturing companies is how best to exploit recent developments in computer-controlled production technology. With its potential to reduce lead times, improve quality and consistency of output, increase throughput and provide greater flexibility in response to changing market demand (Majchrzak, 1988), computer control promises a competitive advantage that cannot be ignored.For some, the ultimate goal of this advanced manufacturing technology (AMT) is the fully computer-integrated, "peopleless," factory of the future (Sharit, Chang, & Salvendy, 1987). At present, however, application is at a much more modest level. A small minority of companies have introduced flexible manufacturing systems and other intermediate levels of computer automation, but, for the vast majority, technical and economic constraints mean that smaller scale applications, such as computer-numerical-control (CNC) machines, are the reality (Corbett, 1987;. CNCs are the building blocks for the eventual development of larger integrated systems, but for now most CNCs will continue to be operated as stand-alone systems involving considerable human support (Hirschhorn, 1986;Valery, 1987).The study reported was conducted at the Social and Applied Psychology Unit of the University of Sheffield, which is jointly funded by the Medical Research Council (MRC) and the Economic and Social Research Council (ESRC) of the United Kingdom.We gratefully acknowledge the cooperation received from the many participants in the study, who for reasons of confidentiality must remain anonymous.