The Council of Coaches’ RRI process comes to an end-Tech Episode 5

For the past 3 years the technical work and design efforts of building the Council of Coaches application has been supported by a hands-on process of keeping the consortium engaged in debating important aspects of the research process in terms of responsibility, ethics and societal interest. This is what we call our RRI-process. From the beginning of the project, the consortium has kept a distinct focus on the following four issues:

Issue #1: Privacy and informed consent

Issue #2: Trust

Issue #3: Handling disagreement between coaches

Issue #4: Keeping healthcare knowledge up to date

As the project is approaching its conclusion, so is our RRI-work. The process has brought us to many hooks and corners of responsibility work and reflection and it has helped the project stay on track in relation to the pinpointed issues. Here some concluding remarks.

Issue #1: Privacy and informed consent

The issue of privacy and informed consent has in many ways been the most remarkable issue we have worked with, in terms of extensiveness as well as urgency. In May 2018 GDPR was implemented throughout Europe, which made reflections and actions within the ethical, social, and legal implications of research not only a timely, but also a mandatory concern. For example, the consortium underwent a steep learning curve in terms of patenting law and legislation and the very different proceedings depending on whether a device is labelled as medical or rather as a life-style hack.

Issue #2: Trust

In designing and building the application, much effort has been invested in toning down the level of trust. This may sound strange, but one of the main purposes of the project has been to enable an application that will encourage users to consider different opinions and different sides of a story – and then make their own decision as to what they are going to do with their health. We want users to adopt a healthy, critical approach, which is why we have taken measures to counteract that the user bestows too much authority on the coaches. Two of the main trust-lowering mechanisms that have been built into the system for these reasons are the cartoony look of the coaches and the amount of humor in the conversational style and exchanges. The humor is intended to create a light atmosphere between the user and the coaches, while their appearance has the purpose of installing a reflective distance to lower the coaches’ authority as they are, after all, ‘just’ cartoony-looking characters, dressed in casual clothes in a living room – not life-like personas dressed in white coats in a virtual doctor’s office or clinic. Still, the aim is not ‘just’ to create critical space and distance. We still want the users to build relationships with the coaches over time, but we want the character of the relation to be ‘just right’.

Issue #3: Handling disagreement between coaches

Regarding the handling of disagreement between coaches, the original idea took into account that we could end up with a very extensive logic-based knowledge base-structure that could automatically come up with arguments or statements for the coaches and then you would have a higher risk for conflicts if you would put new information in. But at least for the Functional Demonstrator that risk has turned out to be significantly lower since we use a lot of scripted dialogue so there is much structure between statements for the coaches. However, this does not mean that the project has changed the intention of enabling situations where the user can choose between the coaches’ perspectives and pick advice from one coach while another says something different. Rather, this scenario is closely tied to a future for the application where artificial intelligence is much more integrated than is the case at this point.

Issue #4: Keeping healthcare knowledge up to date

And finally, regarding the challenge of keeping knowledge up to date, the issue quieted much down during the first couple of years of project work. Part of the reason was that, in the case of the Functional Demonstrator, National Guidelines were implemented. These guidelines are of a nature that has already been discussed by a multidisciplinary team and has been implemented into regular healthcare. This development trumps the multiple sources approach, which went into the original framing of the issue in the RRI Vision Workshop. (This holds true at least until something becomes so complicated that guidelines clash.) At this point in time and at this level of automation, the scenario of the coaches automatically updating medical domain knowledge is not yet relevant, but when we advance and get there, our systems are in principle already designed to be able to handle that.

You can read the full story in the two upcoming deliverables on our RRI process in the COUCH project, D2.7 Responsibility Issues in the Council of Coaches Project: Report on ethical, social, and legal considerations and implications; and D2.8: COUCH Participatory Innovation Process:
A socio-technical integration report on the approach, results and lessons learned from a collaborative management method for responsible innovation

One thought on “The Council of Coaches’ RRI process comes to an end-Tech Episode 5”

Leave a Reply

Your email address will not be published. Required fields are marked *