Call for Action: Why AI in Ergonomics Does not Translate to Improved Work Postures

Abstract

Lately, the integration of Artificial Intelligence (AI) in ergonomic assessment has shown promise in accurately evaluating human posture and motion with regard to ergonomic risk. As AI becomes more ubiquitous, these techniques aim to substitute observational techniques and make ergonomic risk assessment more accessible. However, current AI methods in ergonomics still require extensive human post-diagnosis. As workers lack expert knowledge to interpret ergonomic evaluation, this approach does not benefit them in improving their work conditions. To bridge this gap, we identified the challenges that have to be addressed to make AI solutions more worker-centric. We propose a novel design blueprint, the EPA Loop, aimed at improving usability and include recommendations on how to design future AI solutions. To truly improve work conditions, we argue that AI-based ergonomics must evolve from mere diagnostics to holistic solutions that directly target workers, fostering their understanding of ergonomic risks and providing actionable guidance on how to improve their work posture.