Theoretical approaches to assessing the impact of artificial intelligence on the economy of industry as a basis for selecting indicators for statistical analysis and modelling

Abstract

In the context of the rapid adoption of artificial intelligence (AI) technologies in industry, assessing their impact on all components of the economic system, including output, employment, economic efficiency, and related indicators, has become increasingly important. A wide range of theoretical and practical approaches has been developed to evaluate this impact, including those based on mathematical modeling. However, the scientific literature does not always derive mathematical models from explicit theoretical foundations, while theoretical approaches are not always developed to the level of formal models. As a result, two important questions remain insufficiently addressed: «What theoretical rationale underlies the selection of specific indicators for statistical analysis and their inclusion in a model?» and, conversely, «If AI affects the economy in accordance with particular theoretical principles, which indicators can be used to assess this impact statistically?»

This article analyzes the theoretical foundations of the impact of AI as a closing technology on the economy of industry in order to justify the selection of indicators for subsequent statistical analysis and mathematical modeling. The proposed approach is based on the premise that the theoretical substantiation of the mechanisms through which AI affects the economy, the statistical analysis of relevant indicators, and the construction of mathematical models are complementary tasks. Specifically, the choice of a theoretical framework determines both the set of factors to be tested empirically and the way in which the AI factor should be incorporated into mathematical models, particularly production function models.

In this study, AI is considered primarily as a closing technology, with emphasis placed on its ability to enhance production efficiency without creating fundamentally new products or markets. This interpretation of AI is most closely aligned with the task-based approach, the capital deepening perspective, and the view of AI as a form of capital, as well as with institutional and political economy theories. When the key propositions of these approaches are integrated, AI's impact can be described as follows: at the level of individual work tasks, AI increases productivity and transforms the structure of employment; at the firm level, it functions as a new factor of production, improving efficiency and reducing costs; at the industry level, it accelerates innovation adoption and structural transformation; and at the macroeconomic level, it influences economic growth, labor market dynamics, and the emergence of new institutional arrangements.

For each theoretical approach, the expected effects on the key production factors included in the production function — labor, capital, and AI as a separate production factor — were identified. In addition, general production function specifications were proposed for application in cases where the interpretation of AI's impact on the economy of industry associated with a particular theoretical approach corresponds to the dominant factors driving the dynamics of statistically observable indicators. This provides a basis for the subsequent empirical testing of competing theoretical approaches using statistical data and for determining the most appropriate model specification and method of incorporating production factors into the production function.