Send message to Authors

Paper short abstract

This paper examines how ‘intelligent algorithmic machines’ work in dementia care through the analytical lens of repair work. Drawing on an ethnographic study, the analysis shows how care workers sustain an AI-based telemonitoring system through practices of calibration, workarounds and coordination.

Paper long abstract

Research in STS has increasingly emphasised repair as a generative practice through which sociotechnical systems are maintained (Jackson, 2014). Recent studies have highlighted the human labour required to sustain algorithmic systems (Bailey et al., 2020) and their intrinsic fragility, which becomes visible when these systems fail (Pink et al., 2018). As a result, these systems rely on continuous practices of repair to remain compatible with situated clinical practices (Schwennesen, 2019). Building on these perspectives, this paper proposes repair work as an analytical lens to examine how ‘intelligent algorithmic machines’ operate in dementia care settings.

Drawing on an ethnographic study part of the broader research project ANTICIPATE, the paper examines the invisible work required to align an AI-based telemonitoring system, Ancelia, with nighttime care practices in a dementia unit. The analysis identifies three forms of repair. First, repair occurs when digital infrastructure fails, requiring professional caregivers to replace the telemonitoring system with embodied vigilance and analogue corridor monitoring. Second, repair involves continuously adjusting alarm parameters to reduce false triggers and better reflect residents’ autonomy. Third, repair arises when care workers adapt the prescribed use of technology, such as converting a mobile tablet into a fixed workstation to meet the material demands of bodily care. Rather than simply restoring functionality, repair emerges as the process through which care workers articulate automated monitoring with organisational routines and the embodied needs of residents. Focusing on repair reveals that artificial intelligence in care is sustained through ongoing practices of calibration, workarounds, and coordination.