Can Merging a Capability Approach with Effectual Processes Help Us Define a Permissible Action Range for AI Robotics Entrepreneurship?

Yuko Kamishima,Bart Gremmen,Hikari Akizawa

doi:10.1007/s40926-017-0059-9

Yuko Kamishima, Bart Gremmen + Show 1 more

Open Access

https://doi.org/10.1007/s40926-017-0059-9

Copy DOI

Abstract

In this paper, we first enumerate the problems that humans might face with a new type of technology such as robots with artificial intelligence (AI robots). Robotics entrepreneurs are calling for discussions about goals and values because AI robots, which are potentially more intelligent than humans, can no longer be fully understood and controlled by humans. AI robots could even develop into ethically “bad” agents and become very harmful. We consider these discussions as part of a process of developing responsible innovations in AI robotics in order to prevent catastrophic risks on a global scale. To deal with these issues, we propose the capability-effectual approach, drawing on two bodies of research: the capability approach from ethics, and the effectual process model from entrepreneurship research. The capability approach provides central human capabilities, guiding the effectual process through individual goals and aspirations in the collaborative design process of stakeholders. More precisely, by assuming and understanding correspondences between goals, purposes, desires, and aspirations in the languages of different disciplines, the capability-effectual approach clarifies both how a capability list working globally could affect the aspirations and end-goals of individuals, and how local aspirations and end-goals could either energise or limit effectual processes. Theoretically, the capability-effectual approach links the collaboration of stakeholders and the design process in responsible innovation research. Practically, this approach could potentially contribute to the robust development of AI robots by providing robotics entrepreneurs with a tool for establishing a permissible action range within which to develop AI robotics.

Highlights

In this paper, we first enumerate the problems that humans might face with a new type of technology such as robots with artificial intelligence (AI robots)
How should we deal with them? Should we let Artificial Intelligence (AI) robots steer our lives in whatever direction they will? Or should we develop AI robots that think like a human engineer concerned about ethics? We have reformulated this last question into the main objective of our paper: to define a permissible action range for AI robotics entrepreneurship
This article is structured as follows: first, we briefly examine a central problematic of robotics, that is whether we should control this innovative technology by assigning purposes and ethics for responsible innovation; we introduce the capability approach arguing that society will be minimally good if we share values concerning central human capabilities; we explain the effectual process by which stakeholders collaboratively formulate new purposes and plans, and distinguish it from the causation process, which is controlling or optimising a plan with a purpose

Summary

Introduction

We first enumerate the problems that humans might face with a new type of technology such as robots with artificial intelligence (AI robots). The capability approach offers a perspective on how we can look at and deal with the various goals and values in global societies, while the effectual process model offers a perspective on collaborative design processes that stakeholders might employ in robotics entrepreneurship.

Results

Conclusion