Disrupting and improving work behaviour using machine learning within an IoT environment
Smarthomes are becoming more common, with an IoT environment catering to our needs in exchange for collecting data about our behaviour. In times where burnout is affecting an increasing fraction of white-collar workers, this project explores a speculative scenario in which companies detect behavioural patterns of employees in order to sustain work efficiency and mental health among their workers by providing fully-equipped smarthomes, set up in a modern company town.
A speculative design scenario
In economically advanced countries, the modern employee deals with chronobiological and social disruptions caused by non-standard working schedules and chronic work stress, leading to an increased risk of burnout. Each individual is influenced by various factors to a different extent (ranging from work to social, built and natural environments), uses different coping mechanisms and experiences different impacts on behaviour and health.
This project explores how big tech-companies might exploit the collection of data and potential of machine learning in order to increase work efficiency as well as prevent employees from experiencing mental health disorders by changing the immediate living environment in response to the onset of exhaustion or work stress, thus blurring the boundaries between work and private life. The scenario is set in a speculative modern company town, including company-owned smart-homes and recreation centres.
Detect, analyse, interpret, respond
The modern employee is being influenced by their environment, own personal qualities, job satisfaction, worklife, social network and non-standard work schedules. This has effects on their behaviour, health and therefore on their work performance. In the case of burnout, symptoms such as emotional exhaustion, cynicism and professional inefficiency make themselves apparent. What if an IoT environment interpret one's mental health condition? What if an algorithm is able to detect deviations in one's work performance, analyses possible causes and the smarthome responds according to the interpretation of the system? What if your lights turn off because your home wants you to go to sleep in order to prevent you from overworking yourself?
Focus on posture, shakiness and typing behaviour
This project focuses on three inputs: posture for exhaustion, shakiness for stress and typing behaviour for distraction. While a self-developed touch-sensitive surface (conductive ink and touchsensor) detect one's arm position, an accelerometer attached to a chair measures shakiness or knee jiggling. A type editor – developed in openFrameworks – analyses typing velocity and frequency of hitting 'backspace'.
Although posture could also indicate type of work, this project focuses on the second row of illustrations.
Visualisation and prompts
The wave visualisation changes according to the user inputs: 1) shakiness leads to rapid fluctuation of the wave amplitude, 2) hand positioning changes its overall amplitude, e.g. user exhaustion is communicated by low amplitude, and 3) type velocity influences the sinewave length while error rate results in wave iterations. Prompts appear underneath, e.g. 'Your battery is low; Time for a power nap' in case of exhaustion and distraction.