Governance and legal framework for artificial intelligence management

Artificial intelligence can redefine decision making in companies and organisations. Among other things, it can take advantage of the extensive data concerning competitors, customers and the like. However, there is growing concern that the fairness and transparency of AI is limited, as this technology tends to reproduce existing distortions in the data and thus implicitly to make discriminatory predictions. These limitations hamper the widespread use of AI.

The aim of this project was to develop legal and regulatory frameworks which, on the one hand, ensure a development and use of AI in which equity principles are the core of the technology and, on the other hand, warrant continuous and consistent responsibility for the technology. The interdisciplinary approach of the project team, led by Klaus Möller, compromised IT, management and law.

At the start of the research project in 2020, there were more AI companies per citizen in Switzerland than in any other country. AI «made in Switzerland» is characterised by the principles of democracy and fairness inherent in the Swiss social context, and therefore represents a unique competitive advantage for our country.

The project team analysed existing laws and developped regulatory frameworks that protect the rights of individuals and companies, while enabling the development and use of AI. The researchers studied the mechanisms that promote trust in AI, and designed a management framework that guarantees the governance of AI in organisations. Based on these investigations, they developed design principles that inherently ensure the fairness, accountability and transparency of AI.

The project advanced research into the applicability and efficiency of fair algorithms, shedding light on both the willingness and the limitations surrounding their deployment in public spaces. In parallel, the researchers developed evidence-based recommendations for the governance and oversight of AI within organisational contexts. A set of practical tools and guidelines was also created to support decision-makers and administrative staff in evaluating and managing the use of technologies for automated decision-making.

Three Main Messages

• Integrating fairness in AI positively improves stakeholders’ perception of AI.
• Both technical (fair AI) and societal (governance, regulation, goal transparency) are necessary to increase trust in AI.
• Stakeholders perceive fairness differently depending on whether a human or AI is making decisions.

Governance and legal framework for managing artificial intelligence (AI)