Trustworthy AI and Corporate Governance: The EU’s Ethics Guidelines for Trustworthy Artificial Intelligence from a Company Law Perspective

Abstract

AI will change many aspects of the world we live in, including the way corporations are governed. Many efficiencies and improvements are likely, but there are also potential dangers, including the threat of harmful impacts on third parties, discriminatory practices, data and privacy breaches, fraudulent practices and even ‘rogue AI’. To address these dangers, the EU published ‘The Expert Group’s Policy and Investment Recommendations for Trustworthy AI’ (the Guidelines). The Guidelines produce seven principles from its four foundational pillars of respect for human autonomy, prevention of harm, fairness, and explicability. If implemented by business, the impact on corporate governance will be substantial. Fundamental questions at the intersection of ethics and law are considered, but because the Guidelines only address the former without (much) reference to the latter, their practical application is challenging for business. Further, while they promote many positive corporate governance principles—including a stakeholder-oriented (‘human-centric’) corporate purpose and diversity, non-discrimination, and fairness—it is clear that their general nature leaves many questions and concerns unanswered. In this paper we examine the potential significance and impact of the Guidelines on selected corporate law and governance issues. We conclude that more specificity is needed in relation to how the principles therein will harmonise with company law rules and governance principles. However, despite their imperfections, until harder legislative instruments emerge, the Guidelines provide a useful starting point for directing businesses towards establishing trustworthy AI.

Artificial Intelligence and Sustainable Decisions

“…88 Petrin (2019). 89 Kolbjørnsrud et al (2016); Hickman and Petrin (2020), p 10. 90 Hilb (2020), p 852.…”

Defining organizational AI governance

“…We note that this positioning considers AI governance as governance of AI systems as IT systems. The recent argument that governance may be increasingly conducted by AI [38] should, however, be explored in future research. The relationships between AI governance and the other governance areas are explored in more detail below.…”

From AI ethics principles to data science practice: a reflection and a gap analysis based on recent frameworks and practical experience

“…For managers, whereas the guidelines promote positive corporate governance principles such as stakeholderism, it is yet unclear how to translate them into practice, leaving much room for corporations to use AI to further predominantly or only shareholder-related interests [28]. Moreover, AI as an organizational capability requires numerous (less tangible) human and organizational resources such as interdepartmental coordination, organizational chance capacity, and risk proclivity [42], which are by definition prerequisites for the implementation of any human-centric capability within an organization.…”