Machine Ethics, Allostery and Philosophical Anti-Dualism: Will AI Ever Make Ethically Autonomous Decisions?

Abstract

Essentially, the area of ​​research into the ethics of artificial intelligence is divided into two main areas. One part deals with creating and applying ethical rules and standards. This area formulates recommendations that should respect fundamental rights, applicable regulations and the main principles and values, ensuring the ethical purpose of AI while ensuring its technical robustness and reliability. The second strand of research into AI ethics addresses the question of whether and how robots and AI platforms can behave ethically autonomously. The question of whether ethics can be “algorithized” depends on how AI developers understand ethics and on the adequacy of their understanding of ethical issues and methodological challenges in this area. There are four basic problem areas that developers of machines and platforms containing advanced AI algorithms are confronted with – lack of ethical knowledge, pluralism of ethical methods, cases of ethical dilemmas, and machine distortion. Knowledge of these and similar problems can help programmers and researchers avoid pitfalls and build better moral machines. Unfortunately, discussions in areas that should help to research the field of AI ethics, such as philosophy of mind or general ethics, are now hopelessly infused with a number of autotelic philosophical distinctions and thought experiments. When asked whether machines could become fully ethically autonomous in the near future, most philosophers and ethicists answer that they could not, because AI has no free will and is unable to realize phenomenal consciousness. Therefore, the main proposition of this text is that questions about the ethics of autonomous intelligent systems and AI platforms evolving over time through learning from data (Machine Ethics) cannot be answered by the concepts and thought experiments of the philosophy of mind and general ethics. These instruments are closed to the possibility of empirical falsification, use special sci-fi tools, are based on faulty analogies, transfer the burden of proof to the counterparty without justification, and usually end in an epistemological fiasco. Therefore, they do not bring any added value. Finally, let us stop analysing and overcoming these infertile philosophical distinctions and leave them at their own mercy.