Lay summary

Portrait / project description

AI could redefine decision-making in companies and organizations. Organizations can take advantage of the massive data they have collected on their competitors, customers, etc. Yet, there is growing concern that the fairness and transparency of AI are limited: it tends to reproduce existing biases, present in the data, and therefore makes implicitly discriminatory predictions. These limitations impede the widespread deployment of AI.

This project aims to develop legal and governance frameworks that ensure, on the one hand, the development and use of AI with equality principles at the core of the technology and, on the other hand, guidelines that ensure that responsibility for the technology is constantly and consistently assured.

Background / Initial situation

While AI is not new, it is currently undergoing a rapid development in various fields of research. Our interdisciplinary approach with researchers and practitioners from IT, Management, and Law will help to create a congruent terminology between these different fields.

Switzerland has more AI companies per citizen than any other country, which represents an important economic driver. AI 'made in Switzerland' is shaped by the principles of democracy and fairness inherent in the Swiss social context. It represents a unique competitive advantage for Switzerland.

Objectives / goals

We analyze current laws and develop regulatory frameworks, which protect the rights of individuals and companies (such as fairness and transparency) while enabling the development and deployment of AI. We design and validate a management framework that ensures the governance of AI in organizations. We study the mechanisms that promote trust in AI. We use these results to develop design principles that ensure fairness, accountability, and transparency of AI 'by design'.

Meaning / importance / practical utility

Abstract

Artificial intelligence (AI) has the potential to redefine decision-making in businesses and organizations. Yet there is a growing concern that fairness, accountability, and transparency in artificial intelligence is inherently limited. This results in new governance and legal challenges, and, as a remedy, this interdisciplinary project aims at developing a governance and legal framework for artificial intelligence. To achieve this, we study - and solve - the challenges across a micro/meso/macro-level.
(1) Based on a legal analysis and interdisciplinary analysis, we analyze the current laws and develop regulatory implications, that safeguard individual and companies’ rights (such as fairness and transparency) while also permitting the development and deployment of AI.
(2) For organizational governance, we develop and validate a management framework that ensures AI governance in business and organizations.
(3) For human perception and fairness, we study mechanisms that foster trust in AI.
For (4) technological design principles: here we use the previous findings to develop design principles that yield fairness, accountability, and transparency in AI “by design”.

In keeping with this, we advance “AI made in Switzerland” towards democratic elements. Across all work packages, we yield actionable interdisciplinary outcomes that aid the implementation of AI in practice.

Last updated:23.08.2023