From the Home to the Factory and Back: A History of Smart Homes, Machine Learning Systems, and Labor Discipline

Over the past two decades, algorithmic management systems have been integrated into labor processes across many industries. Wristbands track workers’ productivity in warehouses. AI-infused cameras purportedly measure the attentiveness of Amazon drivers. “Emotion AI” systems scrutinize workers’ facial expressions and voice tones to infer their emotional states. While companies claim they deploy these devices to ensure their workers’ safety and well-being, this rhetoric fails to obscure the coercive, invasive, and exploitative aspects of digital Taylorism.

One may think that the techniques of algorithmic management were developed in the factory. As this paper demonstrates, several key tools underlying today’s datafied workplace surveillance were developed, tested, and refined within smart home projects of the late twentieth century. The paper examines two prototypes constructed between 1980-2000 – Ken Sakamura’s “Tron Intelligent House” and Michael Mozer’s “Adaptive House”. Both Sakamura and Mozer engineered systems to automate household activities through the use of computing technologies and techniques, such as sensors and machine-learning models. Thus, the designers programmed these systems to minimize certain kinds of domestic labor. Ironically enough, many of their techniques have since been repurposed to maximize the productivity of workers.

Of course, we all know the same tool can be applied toward conflicting goals, and this history is not meant to provide a reminder of this banal fact. By examining the connections between the smart home and algorithmic management, this paper highlights the remarkably similar logic and rhetorics shared between these two very different historical and sociotechnical arrangements. Both centrally entail the devaluation of certain kinds of labor by automation. And most importantly, discourses of health and safety played a crucial role in distracting from the devaluations entailed by each system.