Addressing AI ethics through codification

Abstract

AI ethics rapidly becomes one of the most significant issues in assessing the impact of AI on social welfare and development. A technology that does not meet the ethical criteria of a society is likely to face a long and hard process of acceptance regardless of its potentially tremendous positive potential for long-term socio-economic development. The development of artificial intelligence (AI) technologies is undoubtedly associated with the need to answer ethical questions, and the perception of AI in society will be largely determined by compliance with ethical criteria, whether written or not. At the same time, AI as a technological system itself does not have a natural ethical content; the authors believe that in practice ethical concerns may be addressed by means of ethical codes and compliance rules that articulate what constitutes ethical behaviour in specific areas of application of AI systems. Such a set of rules (a code for AI ethics) could be followed by all actors throughout the complete lifecycle of the system starting with the design stage. The specification of general ethical principles as industry-specific codes of practice would also facilitate classification, evaluation and measurement of systems, both at the technical level and at the level of public perception and trust.The article considers examples of codification of ethical principles and offers several approaches for practical use in solving issues of ethics in the field of AI at the national and international level.