The Ethics of Artificial Intelligence: From Philosoph-ical Critique to Practical Responsibility

This paper is devoted to analysing the ethical challenges associated with the development of artificial intelligence (AI) and finding ways to overcome them through the synthesis of philosophical reflection, engineering methods and collective responsibility. The contradictions between the autonomy of technologies described by Jacques Ellul and the need for them to be subject to ethical principles are explored by analysing the views of philosophers (Capurro, Borgmann, Verbeek, Magnani) on human interaction with digital systems. The methodology includes a the-oretical analysis of philosophical concepts and a qualitative review of practical cases such as the biased COMPAS algorithm and Google Health’s transparent diagnostic systems. The paper proposes tools for minimising AI risks: a three-layer audit (legal, technical, ethical), Mepham’s ethical matrix and the ‘three lines of defence’ model. Special attention is given to the role of ethical auditing as a mechanism that combines normative principles (OECD AI, IEEE) with transparency practices (XAI, Fairlearn). It is concluded that sustainable AI implementation requires an interdisci-plinary approach, integration of philosophical critique into engineering design and global standardisation of auditing. Consequently, AI ethics should not be a barrier but a foundation for innovation, balancing technological progress with the preservation of human-centred values.