THE PHENOMENON OF ARTIFICIAL EVIL (ON THE RISKS OF MORAL ‘QUASI-DEPRAVITY’ IN THE DEVELOPMENT OF ARTIFICIAL INTELLIGENCE)

The article considers the phenomenon of artificial evil, expressed in the manifestation of properties by artificial intelligence similar to human moral depravity. The author discusses possible risks of moral quasi- depravity (similarity- depravity) in the evolution of artificial intelligence, in connection with the complication and expansion of its functions to the level of GPT-4 and GPT-5 chatbots. The moral quasi-depravity of AI is erroneous or misleading answers and solutions of AI that deviate from the specified parameters towards biased goals and outwardly resemble human vices. As such AI vices, researchers usually highlight: deception and generation of false information, lies out of philanthropy, theft of other people's information (AI plagiarism), and moral unscrupulousness of AI. In this article, the main subject of consideration is the phenomenon of deception and false information on the part of AI. The article emphasizes that this is not about the freedom to choose a vicious decision based on conscious intent, but about the effect of Algorithmic Bias and the lack of AI alignment. According to the author, such moral deviations will increase with the further development and complexity of AI systems and will increasingly resemble the phenomenon of human depravity, reaching the limit of possible freedom to choose negative decisions (quasi-free will of AI).