The Ethics of AI Ethics: An Evaluation of Guidelines

  • PDF / 817,149 Bytes
  • 22 Pages / 439.37 x 666.142 pts Page_size
  • 71 Downloads / 251 Views

DOWNLOAD

REPORT


The Ethics of AI Ethics: An Evaluation of Guidelines Thilo Hagendorff1  Received: 1 October 2019 / Accepted: 21 January 2020 © The Author(s) 2020

Abstract Current advances in research, development and application of artificial intelligence (AI) systems have yielded a far-reaching discourse on AI ethics. In consequence, a number of ethics guidelines have been released in recent years. These guidelines comprise normative principles and recommendations aimed to harness the “disruptive” potentials of new AI technologies. Designed as a semi-systematic evaluation, this paper analyzes and compares 22 guidelines, highlighting overlaps but also omissions. As a result, I give a detailed overview of the field of AI ethics. Finally, I also examine to what extent the respective ethical principles and values are implemented in the practice of research, development and application of AI systems—and how the effectiveness in the demands of AI ethics can be improved. Keywords  Artificial intelligence · Machine learning · Ethics · Guidelines · Implementation

1 Introduction The current AI boom is accompanied by constant calls for applied ethics, which are meant to harness the “disruptive” potentials of new AI technologies. As a result, a whole body of ethical guidelines has been developed in recent years collecting principles, which technology developers should adhere to as far as possible. However, the critical question arises: Do those ethical guidelines have an actual impact on human decision-making in the field of AI and machine learning? The short answer is: No, most often not. This paper analyzes 22 of the major AI ethics guidelines and issues recommendations on how to overcome the relative ineffectiveness of these guidelines. AI ethics—or ethics in general—lacks mechanisms to reinforce its own normative claims. Of course, the enforcement of ethical principles may involve * Thilo Hagendorff thilo.hagendorff@uni‑tuebingen.de 1



Cluster of Excellence “Machine Learning: New Perspectives for Science”, University of Tuebingen, Tübingen, Germany

13

Vol.:(0123456789)

T. Hagendorff

reputational losses in the case of misconduct, or restrictions on memberships in certain professional bodies. Yet altogether, these mechanisms are rather weak and pose no eminent threat. Researchers, politicians, consultants, managers and activists have to deal with this essential weakness of ethics. However, it is also a reason why ethics is so appealing to many AI companies and institutions. When companies or research institutes formulate their own ethical guidelines, regularly incorporate ethical considerations into their public relations work, or adopt ethically motivated “self-commitments”, efforts to create a truly binding legal framework are continuously discouraged. Ethics guidelines of the AI industry serve to suggest to legislators that internal self-governance in science and industry is sufficient, and that no specific laws are necessary to mitigate possible technological risks and to eliminate scenarios of abuse (Calo 2017). And even whe