Work-related accidents and illnesses cost EU €476 billion a year, and presenteeism (coming to work, but not focusing on work) is at least as costly as illness-related absenteeism or disability . Due to demographic change, EU countries increase retirement age, but many workers leave labour market rather early. In USA, job stress is estimated to cost $300 billion a year in employee turnover, absenteeism, diminished productivity and direct medical, legal and insurance fees.
Presenteeism, absenteeism, employee turnover and early retirement often originate from work-related risk factors, e.g., job stress and organizational culture. Thus solutions to mitigate workplace-related risk factors have a huge market potential, currently practically unexplored.
Mental health in the workplace is defined as a positive state of psychological well-being and not only as the absence of mental health disorders , and this project focuses on detection and mitigation of poor mental health conditions, such as stress and burnout, which may have not yet resulted in a diagnosed mental health disorder.
As sensors and IoT technologies are becoming abundant, they can serve as a basis for providing personalised services to improve work satisfaction/ productivity and life quality/ health of employees. Currently, however, research into assessing mental conditions is mainly conducted in the labs. Results cannot be directly used in real life because the most harmful stress type, long-lasting stress, cannot be induced in lab studies. Another problem is that real life requires AI methods, highly adaptive to differences in user requirements, privacy requirements and availability of data sources in different workplaces, as well as changes in workers' tasks and locations, inevitable in long term. Majority of existing solutions, however, rely on supervised learning methods, requiring too large sets of labelled data for each situation to enable such adaptivity. In addition, existing solutions to assess mental conditions are based on not-so-affordable, not-so-accurate and inconvenient for long-term everyday use wearable devices, whereas research into use of environmental sensors for this purpose is just emerging.
Mad@Work project aims at major breakthrough in development of software intense applications that combine multiple heterogeneous environmental and/ or wearable data sources into actionable information for improving employees' wellbeing, engagement and performance. Mad@Work project will develop truly unobtrusive, privacy-safe, appealing solutions, smoothly integrated into work environment and appropriate for long-term use in diverse real life settings:
• methods to assess mental conditions, based on various environmental sensors in the workplace, smart phones and/ or wearables and AI algorithms, requiring little or no human supervision for adaptation to different conditions
• methods to help individuals to improve own mental wellbeing and productivity
• methods to enable collective awareness and support towards empathic work culture
• methods to aggregate and visualise data of multiple persons into “organisational barometer” to compare different teams, environments, office layouts etc.
• extension of the IoT standards with relevant AI and mental health attributes.
In addition to the above-listed innovation, Mad@Work aims at semantic meta-modelling of engaging vs. poor working environments and human conditions and related privacy and ethical safeguards, that would allow to use sensor data in Business Processes. Those models will be transformed to get SaaS application supporting those Business Processes.
