AI AGENTS: THE END OF ECONOMIC THEORY OR A NEW BEGINNING
Abstract
The aim of the article is to develop the theoretical provisions of economic theory concerning demand associated with partially autonomous decision‑making by AI agents in markets. The paper provides an overview of the approaches taken by various schools of economic theory to the issue of market transformation and analyses the trend towards using AI assistants as market agents. This forces a re‑evaluation of traditional ideas about supply, demand, and the market mechanism. The article examines the stages of automation of technical and socio‑economic processes, the transition from the automation of individual operations to systems capable of making decisions autonomously. It analyses the trends in the use of AI intermediaries, the factors involved, and the main directions for the development of the use of partially autonomous agents. The paper also analyses scientific works containing descriptions of the types of agents based on artificial intelligence that are currently in use. A clarification of the definition of market demand is proposed, taking into account the activities of AI agents; the concept of demand elasticity under the influence of intermediary activities of artificial intelligence is presented. The partial autonomy of decisions made by AI assistants, on the one hand, is part of the aggregate demand and is technically based on it; on the other hand, it transforms demand and acts as an active component influencing purchasing decisions. This is a fundamental difference in the economic theory of demand; previously, technical devices could not intervene and shape market processes. The results of a sociological study of the readiness of residents of the city of Barnaul to use AI intermediaries are presented. It has been revealed that the age group 31-60 years is the most receptive to using semi‑autonomous devices as trading agents. Studies have shown that customers value having control over their spending, which needs to be taken into account when developing AI assistants.