The book “Ethics of Artificial Intelligence” has been published in Russia — it is the country’s first textbook that combines classical philosophical ethics with engineering methods for formalizing moral norms in artificial intelligence architectures. The authors of the book are Ekaterina Tikhomirova, a professor at the Department of Philosophy, Ontology and Theory of Knowledge at MEPhI, and Roman Dushkin, a senior lecturer at the Department of Cybernetics at MEPhI.
The issue of responsibility for decisions made by algorithms has become more acute against the backdrop of the accelerated integration of large language models into business processes, medicine, and public administration. At the same time, the number of declarations and “AI ethical codes” is growing, but they are rarely translated into engineering solutions. Experts in the field of technology regulation note a shortage of specialists capable of working simultaneously with the philosophical framework and technical tools of alignment — aligning the behavior of models with human values.
At the same time, Russia is actively developing its regulatory framework in the field of AI, which makes training personnel with such an interdisciplinary profile a practical task rather than a purely academic one. The book was co‑published by the MNIIPU Publishing House (the International Research Institute of Management Problems) and the company “Ecosystem Digital Solutions” (LLC “EDS”).
The preface to the textbook was written by Academician of the Russian Academy of Sciences, Doctor of Economic Sciences Sergei Glazyev; the afterword was written by Alexander Ageev, General Director of MNIIU, Head of the Department of Business Project Management at MEPhI. ECR acted not only as a co‑publisher but also as the developer of the digital support for the manual — in the company’s training center, the book has already been included in the specialist training program.
The publication is based on the differentiation of concepts that are often confused in public discussion: ethics as a science, morality as a system of cultural norms, morality as a personal development of these norms and law as their formalized, institutionally fixed slice.
The authors show that substituting ethics with law leads to technocratic minimalism — a state in which everything that is not directly prohibited by law is considered permissible.
The book pays special attention to the cultural variability of moral categories: the authors examine how universal concepts of goodness and justice acquire different substantive content in Chinese, Japanese, Russian, and Western traditions — and why this is crucial for developing AI systems that operate in different cultural environments.
The engineering part of the textbook traces the history of the formalization of morality from Soviet‑era expert systems to modern methods of working with large language models — including RLHF, Constitutional AI, red teaming, and governance pipeline. Separate sections are devoted to the typology of failures and abuses in AI systems, as well as to the criticism of declarative ethical codes as a tool that creates the illusion of control without real engineering implementation.
“Ethics of Artificial Intelligence” opens the series of publications by MEPhI devoted to the philosophical reflection on technologies. The next books in the series will be “The Aesthetics of Artificial Intelligence,” “The Anthropology of Artificial Intelligence,” and “The Philosophy of Artificial Intelligence” — the authors describe this sequence as an attempt to systematically describe how technological transformations are changing the foundations of culture and ideas about humanity.
The textbook is aimed at students and postgraduate students in IT and philosophy, developers of AI systems, legal scholars, and researchers — anyone working at the intersection of technology and humanities. The manual is recommended as a basic material for the course “Philosophy of Science and AI Ethics” and related disciplines at universities.
Source of information: RBC
Image source: Ruplexity.ru
