Addressing trade-offs in co-designing principles for ethical AI: perspectives from an industry-academia collaboration

Abstract

The development and deployment of artificial intelligence (AI) has rapidly outpaced regulation. As a result, many organizations opt to develop their own principles for the ethical development of AI, though little research has examined the processes through which they are developed. Prior research indicates that these processes involve perceived trade-offs between competing considerations, and primarily between ethical concerns and organizational benefits or technological development. In this paper, we report on a novel, collaborative initiative in Japan between researchers in the humanities and social sciences, and industry actors to co-design organizational AI ethics principles. We analyzed the minutes from 20 meetings from the formative phase of the development of these principles using an inductive process drawing on thematic analysis, to identify the issues of importance to participants. Through this, we identified four core trade-offs faced by participants. We find that, contrary to prior literature, participants were not just concerned with trade-offs between ethical concerns and organizational benefits or technological development, but also between competing, ethically-oriented considerations. We use the results of this study to highlight a need for further research to understand the longer-term impact on organizations and on society of organization-led approaches to AI ethics.