AI ETHICS: FAIRNESS, TRANSPARENCY AND ACCOUNTABILITY CHALLENGES

This article presents a theoretical analysis of the three core principles of AI ethics: fairness, transparency and accountability. It examines racial, gender and social forms of algorithmic bias and its sources at the data, modeling and deployment stages. The article also discusses the “black box” problem, the interpretability of decisions (Explainable AI) and their practical and legal limits. In addition, it justifies the “Ethical by design” principle, which builds ethical requirements into a technology’s life cycle from the very beginning, and its use in education and other high-risk areas. The main conclusion is that AI ethics is not a restriction added after the technology is built, but a basic precondition for a system’s reliability, legality and social acceptability.