Moral Machines; the implications of AI for the practices of care.

Dr Simon Bailey's research centres on the relationship between practices of care and formal organisational conduct. In his talk, he will explore this relationship by focusing on the use of new technologies, such as Artificial Intelligence (AI), within the organisation and delivery of care. 

Practices of care are complex; they do not submit easily to planning or measurement, rather care requires situated attention to the particularities of cases. Care work is a process of finding a fit between these particularities and professional knowledge and practice, within the possibilities and constraints of particular organisational systems. Reconciling the differences and tensions that emerge from this attempt requires the collective exercise of moral agency, which historically has been understood as a matter of professional judgment. The use of AI in care alters this. This is partly because AI is built upon categories and classifications derived from statistical norms of illness and disease which are pre-coded. If particular experiences of illness diverge from these norms then the technology requires the means to interpret the divergence between its pre-coding and the new data. While it is possible for AI to learn and adapt to new data this can only be achieved through the application of rational logic, yet, the judgement required in each new case also contains ‘irrational’ elements – emotion, politics, ethics – which are matters of experience, not logic.  

This leads us to a central concern in the use of AI in care systems: AI can be ‘descriptively’ responsible, and it can cause things to happen independent of human input, yet it cannot be responsible in a ‘normative’ or ‘possessive’ sense. AI can therefore independently execute actions for which humans are accountable. How then can organisations meet the moral imperatives taking shape out of this novel cooperation, in which humans and AI share responsibilities, but on unequal terms?