Extension of Interval-Valued Hesitant Fermatean Fuzzy TOPSIS for Evaluating and Benchmarking of Generative AI Chatbots

Source abstract

To aid in the selection of generative Artificial Intelligence (GAI) chatbots, this paper introduces a fuzzy multi-attribute decision-making framework based on their key features and performance. The proposed framework includes a new modification of the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), adapted for an interval-valued hesitant Fermatean fuzzy (IVHFF) environment. This TOPSIS extension addresses the limitations of classical TOPSIS in handling complex and uncertain data capturing detailed membership degrees and representing hesitation more precisely. The framework is applicable for both static and dynamic evaluations of GAI chatbots in crisp or fuzzy assessments. Results from a practical example demonstrate the effectiveness of the proposed approach for comparing and ranking GAI chatbots. Finally, recommendations are provided for selecting and implementing these conversational agents in various applications.