Secure User-to-Chatbot Authentication Using One-Time Session Keys Based on DID

Artificial Intelligence (AI) Chatbots, traditionally limited to providing predetermined responses, have evolved with advancements in technology to incorporate Robotic Process Automation (RPA), enabling them to deliver automated work systems around the clock.Given that identity authentication is often necessary for these automated processes, AI Chatbots now store sensitive user authentication data, including certificate passwords, certificates, and private keys.This creates a vulnerability, as an attacker infiltrating an AI Chatbot could potentially expose user authentication information, allowing the attacker to forge the user's identity and compromise data.This paper introduces a mutual authentication method between users and AI Chatbots in RPA systems that utilizes a one-time session key based on Decentralized Identifier (DID), eliminating the need to store sensitive user information directly in the Chatbot, thereby ensuring reliable and secure automation.By leveraging DID, users enhance their privacy, and the use of a one-time session key ensures each authentication's independence, enhancing overall security.Additionally, we compare the time efficiency of this DID-based authentication method against the traditional Public Key Infrastructure approach, demonstrating that our proposed method is more efficient during multi-factor authentication processes between users and AI Chatbots.