THE PROBLEM OF AUTHORSHIP AND THE ETHICS OF USING ARTIFICIAL INTELLIGENCE TO CREATE VISUAL WORKS

Abstract

The article examines a complex set of issues related to the authorship and ethics of using generative models of artificial intelligence for creating visual works. It explores the functioning principles of leading platforms such as DALL-E, Midjourney, and Stable Diffusion, which allow generating high-quality images based on textual descriptions. Particular attention is paid to the formation of large-scale datasets used to train these models. It is noted that such datasets, particularly LAION-5B, often contain millions of works by real artists, included without their explicit consent and proper compensation. This raises a number of ethical and legal issues, as artists and photographers may consider such use of their works as a violation of their copyrights for reproduction, distribution, and processing of the works. Other problematic aspects are also highlighted, including the potential devaluation of original creativity in the context of AI technology development and the risks of creating offensive or harmful content by algorithms that reproduce biases present in the training data. It is noted that these issues are the subject of active discussions and legal disputes, which may have far-reaching implications for AI regulation and intellectual property protection. Special attention is given to the psychological aspect of the problem, namely the impact of AI anthropomorphization on the perception of responsibility. Studies show that attributing human qualities to AI can shift the distribution of responsibility between the algorithm and the user, blurring ethical boundaries. The article proposes possible ways to reconcile the interests of various parties, such as creating compensation funds for artists whose works are used in AI, developing standardized licenses, adapting legislation, and implementing technological solutions for ensuring transparency and control. The need for an interdisciplinary dialogue involving artists, developers, lawyers, and other stakeholders is emphasized in order to develop balanced solutions that will allow unlocking the potential of generative AI models in art while respecting the rights and interests of creators and society.