‘Rethinking trust and consent in mobile behaviour change support systems’

One of the most important ethical issues raised by support systems for behaviour change is consent (Davis 2009, Spahn 2012), which is also a bedrock ethical value in health and medicine (Beauchamp 2011). Mobile support systems for behaviour change pose special problems for consent, because they are ubiquitous, data-intensive, and involve artificial intelligence. It is often difficult to know the underlying processes by which technologies for behaviour change influence behaviour, to predict the effects they might have on individuals and society, and to implement effective mechanisms and procedures for monitoring those effects. Technologies for behaviour change collect extensive data with many potential uses, but the exact uses are difficult to describe and evaluate in advance. Because of these difficulties, Barocas and Nissenbaum have gone so far as to describe “the absurdity of believing that notice and consent can fully specify the terms of interaction between data collector and data subject” (2014, 45).

Unlike traditional medical technologies such as pills and implantable devices, we might say that these mobile BCSS are “nebulous”, involving both a physical “cloud” of data and devices, as well as opacity and unpredictability of the underlying mechanisms by which they achieve their effects. Let us call this the nebulousness problem for consent. Normally consent involves providing a person relevant information about what he or she is consenting to, and what possible effects can be expected. The problem with technologies for behaviour change is that it is very difficult to specify for the user what effects the use of the technology will have on them (and by what mechanism), and what possible uses their data may serve in the future.

In the second project, we investigate the prospects of an alternative model of consent more appropriate to the context of mHealth. We describe this as a distributed model of consent, in which individual consent is not expected to do all of the work alone, and more emphasis is placed on participatory processes involving social and political consent by stakeholders, as well as enhancing trustworthiness and changing how we communicate about trustworthiness to users.