AI-rendered Advice and Management Decisions

12 Pages Posted: 4 Dec 2022

See all articles by Dimitrios Linardatos

Dimitrios Linardatos

University of Mannheim; University of Liechtenstein

Date Written: June 16, 2022


Artificial Intelligence (AI) is arguably the cutting-edge technology in the field of process automation, classification, clustering and extraction of in-formation out of large data sets. The current era of digitization and the ap-proaching era of the Internet of Things will create a decision-making envi-ronment that can hardly be managed without the support of AI.

However, AI has often been discredited as opaque and as a black box. It is claimed that no one can assess how AI reaches to its conclusion. In some cases, AI’s behavior does not match the predictions of its creators, and its actions may not be traceable or comprehensible by human users, since they are lacking the ability to give a causal interpretation to the system’s results. Assuming that this criticism was true: Is it reasonable if the corporate’s directors rely on the assessment of an AI-system to make business decisions?

The paper attempts to provide some (innovation-friendly) answers to this question. It focuses on the directors’ fiduciary duties of care in stock companies according to Section 93 of the German Stock Corporation Act (SCA). In that respect, it is addressed whether the principles devised in the so-called ISION ruling of the German Federal Court of Justice from 2011 (II ZR 234/09) are accordingly applicable when the advice is provided by an AI system rather than by a human.

Keywords: management liability, AI, artificial intelligence, corporate law, fiduciary duties of care, black-box-problem

Suggested Citation

Linardatos, Dimitrios, AI-rendered Advice and Management Decisions (June 16, 2022). Available at SSRN: or

Dimitrios Linardatos (Contact Author)

University of Mannheim ( email )

Mannheim, 68163

University of Liechtenstein ( email )

Fuerst Franz Josef-Strasse
Vaduz, 9490

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics