A Privacy-Preserving Framework for Mental Health Chatbots Based on Confidential Computing

摘要

Mental health chatbots can provide psychological counseling services to patients at any time regardless of time and location, which can not only relieve patients’ ailments but also reduce the workload of psychologists. In order to provide patients with accurate diagnosis and treatment services, mental health robots inevitably collect patient-related information during the communication process with patients, which is often very sensitive and must be well protected. There is a lack of targeted research on how mental health chatbots can provide systematic privacy preserving for patients in the process of providing mental health services to them. In this paper, we propose a privacy preserving framework based on blockchain and confidential computing that can provide comprehensive privacy preserving for patients during mental health chatbot services. We conduct tests using existing mental health chatbots, and the experimental results demonstrate that our proposed framework can meet the requirements for privacy preserving and computational performance of mental health chatbots.