Ethical AI: Proposal to bridge the gap in EU regulation on trustworthy AI and to support practical implementation of ethical perspectives

Abstract

In 2020, GPT-3 defined itself as a thinking robot. The history of AI development is identified with machines becoming increasingly intelligent, but behind it lies the human factor, the soaring of the human mind. However, the question of machine ethics is also a question of cultural ethics. Based on in-depth interviews conducted in seven industries, the author reveals that ethical considerations are not yet taken into account in the development of AI systems. To support practical implementation, the author identifies two shortcomings based on a comparative analysis of the EU’s AI Act and Ethical Guidelines for Trustworthy AI: (1) missing ethical sensitisation and training of AI system developers and supervisors; (2) suggested approaches to handling harmful feedback loops and decision-making biases. The author uses the philosophical and ethical heritage of 21 philosophers as a compass to propose solutions for the identified gaps and deficiencies of organisational integration.