Psychological and Technical Drivers of Anthropomorphism in Artificial Intelligence Tools: A Fractal Fuzzy Analysis Across Healthcare, Finance, and Marketing Industries

Abstract Anthropomorphism in artificial intelligence tools is central to human-technology interactions. However, the lack of systematic identification of the most influential factors influencing the development of this phenomenon in the literature presents a significant problem. When the factors that reinforce anthropomorphism are not clarified, both conceptual ambiguities and practical problems arise. Existing studies in the literature emphasize the importance of these factors but generally focus on single factors and do not sufficiently prioritize the combined evaluation and prioritization of different criteria. This deficiency creates a significant research gap. The aim of this study is to identify the most critical factors affecting the development of anthropomorphism in artificial intelligence tools and to reveal the differences between these factors across sectors. To this end, a comprehensive literature review is conducted, and a total of ten criteria are identified under three dimensions. In this study, three separate fuzzy multi-criteria decision-making models are developed for the healthcare, finance, and marketing sectors, and the dataset is generated using the opinions of five experts in the field. To consider the demographic characteristics of the experts, a Euclidean distance-based expert weighting method is applied, and the updated SIWEC technique, which integrates standard deviation and expert weights, is used to calculate the criteria weights. Furthermore, fractal-based Sierpinski Triangle fuzzy sets are adapted to the proposed model to model uncertainty more flexibly and precisely. The main contributions of the proposed model to the literature can be summarized in four dimensions: (1) providing a more accurate method compared to existing triangular, trapezoidal, spherical, and type-2 sets by developing new fractal-based fuzzy sets; (2) enabling the development of sector-specific strategies by conducting comparative analyses across three different sectors; (3) obtaining more realistic criterion weights with the updated SIWEC technique; and (4) using an expert weighting approach that produces more rational results by overcoming the problem of assuming equal weighting of expert opinions. According to the findings, the most important criterion in the health sector is the perception of reliability (.141), in the finance sector, the perception of reliability (.177) and ethical compliance (.172), and in the marketing sector, behavioral consistency (.119) comes to the fore.