后疫情时代青年新冠感染人群的自助心理干预—基于GPT-4构建PST聊天机器人

o assist young people infected with COVID-19 restore and develop a balanced state of mental health after the pandemic, we have developed an online self-help psychological intervention robot that can complement existing mental health resources. First, we utilized prompting engineering techniques to build a chatbot skilled in Problem-Solving Therapy (PST) based on the large language model GPT-4. Then, we conducted pre-testing and formal experiments to verify the effectiveness of the chatbot. The results of the pre-testing indicated that the chatbot followed the core work steps of PST during interactions with users. The results of the formal experiment showed that the PST chatbot performed better than the ordinary chatbot in terms of problem identification and problem-solving dimensions, indicating that the PST chatbot can help users quickly locate the problems that trouble them and develop feasible problem-solving plans. However, there was no difference between the PST chatbot and the ordinary chatbot in terms of relationship quality, and no differences were found in the evaluation of the two chatbots based on gender and post-COVID symptoms. This suggests that the PST chatbot did not significantly improve the quality of human-machine relationships, but the general acceptability and wide applicability of chatbots still have certain advantages in practical applications. The research results support the possibility of using large language models in innovative implementations of psychological self-help interventions.