Responsible Ethical Learning with Robotics

Responsible Ethical Learning with Robotics – the REELER project - is a research-driven, multidisciplinary collaboration between 4 partners:
- Aarhus University, Danish School of Education (DPU) (Denmark),
- Ab.Acus (Italy),
- De Montfort University (UK),
- University of Hohenheim (Germany).

REELER has identified a clear gap between the people funding and developing robots (i.e. robot makers) and the people expected to use the robots (i.e. affected stakeholders). The project has thus explored possibilities for more collaboration between social scientists and humanists (SSH) and developers of robotics technologies, to ensure more responsible technological research and innovation that goes hand in hand with societal needs and expectations.

Robots have the potential to radically alter human societies as they are expected to increasingly co-exist with or replace humans. While there is a growing amount of literature predicting the economic and social impacts of robots, we still lack knowledge of how robot makers’ visions of a robotic future may differ from affected stakeholders’ needs and concerns, and how these ideas come through in roboticists’ design practices.

REELER has focused on everyday practices in robot development, the collaborations that make development possible, and the learning that occurs (or doesn’t occur) in these processes. Through collaborative learning and 11 ethnographic multi-sited case studies of robots in Europe, our aim has been to raise awareness of the need for increased human proximity in robotics development. Human proximity refers to an alignment with researched human needs and values, which are relevant from both ethical and market standpoints.

REELER has found that, in general, robot developers need to work in closer collaboration with an expanded group of affected stakeholders to make responsible robots. By expanded group of affected stakeholders, we mean people who are directly or indirectly affected by the robotics technology, but also those who are distantly affected – i.e. stakeholders whom may never see or touch the robot, but still feel the effects of it. The REELER roadmap at https://responsiblerobotics.eu/ offers proactive steps to raise awareness of this human potential in robotics development, with special attention to distributed development, relational responsibility, ethical and societal issues, collaborative learning, and the economic and societal impacts of robotization. Our findings are summarized in the publication Perspectives on Robots with ethical guidelines for Human Proximity in robot development. Here we also present our main policy recommendation, which is the introduction of Alignment Experts in future robotics projects. Alignment Experts can help to include the voices of these new types of affected stakeholders in robot development.

Being a multidisciplinary research collaboration, REELER initially focused on establishing a solid foundation for common ground and collaborative learning within the project. This was achieved through literature reviews of relevant concepts and their implication for our research practices, a mapping of selected robotics activities per country in Europe, and a Best Practice Research and Observation Guide.

Through our ‘multi-variation’ approach, we have explored 11 different robot types in multi-sited fieldworks in 13 different European countries. This has ensured variation across country/nationality, human proximity, robot type, sector and/or application, and organization and funding type. In all, we have conducted 160 qualitative interviews. Each case study is summarized in a report providing a review of the type of robot and an overview of empirical findings.

After the fieldwork period, our data processing and analytical efforts began. To look for variation across cases is also a methodological take in our work, where we seek to identify diverse or aligned understandings of robotics among the interviewed robot makers and affected stakeholders. We identified 12 analytical themes that all feed into one of our main outcomes: the collectively written research publication Perspectives on Robots.
Moreover, REELER has developed and tested two tools for collaborative learning across disciplines and between affected stakeholders-robot makers. These tools are Mini-publics and Sociodrama. Mini-publics have proven to be a strong knowledge-sharing tool to facilitate collaborative learning, and it is our belief that it can improve opportunities for citizens to contribute to parliamentary deliberation on a given topic. https://responsiblerobotics.eu/ offers guidelines for running mini-publics in project contexts.
Sociodrama and social drama have also rendered positive results of facilitating interdisciplinary collaboration that allow for new perspectives on robot design and robot uptake to be explored. Nonetheless, these action methods do come with certain limitations and thus cannot be defined as easily-accessible tools.

Alongside this work, REELER has developed a computer model on the economic impact of the introduction and diffusion of robotics notably in terms of employment, composition of the labour force, wages, income inequality, and need for labour mobility (i.e. upgrading skills). This includes estimates of various scenarios of economic added value as well as potential negative effects. It is used to study the impact of various policy measures such as robot tax, universal basic income, and flexicurity. The team has made great efforts to integrate innovation economics perspectives with the qualitative ethnographic data, which ultimately translated into an ABM-inspired ‘serious game’ which proved useful for clarifying how agents act in relation to marketing and design.

Our main results, presented at our REELER roadmap website (https://responsiblerobotics.eu/) are:
- The Human Proximity Model; a descriptive and prescriptive model for collaboration between robot makers and affected stakeholders
- Perspectives on Robots; an interdisciplinary publication that challenges existing notions of robots and users
- Online Toolbox; an interactive website presenting core results through fun and interactive games and illustrations
- BuildBot; a board game using data from ethnographic interviews to simulate a reflective robot design process
- Adapted Mini-Publics; new methods for engaging with different stakeholder groups

REELER researchers have disseminated results in a range of scientific journals and at various conferences; always actively involving the audience in activities of collaborative learning to engage and influence robot makers and SSH-researchers with REELER knowledge.
Conferences include
● ERF, 2017-2019
● 4S, 2017
● RoboPhilosophy 2018
● Responsible Robotics, 2018
● HRI, 2018-2019.
● SIENNA project, 2019.
● ICRES, 2019
● European Research and Innovation Days, 2019
● HAI, 2019
● IROS, 2019
● AAA, 2019

REELER has progressed beyond the state of the art by expanding the concept of ‘users’ to also include indirectly and distantly affected stakeholders as a relevant target group for future robots.
Our ethnographic work and outreach activities has had an impact on the robotics community and civil society by raising awareness of the social dimension of robotics.

Periodic Reporting for period 2 - REELER (Responsible Ethical Learning with Robotics)

Diese Seite teilen

Herunterladen