應用聊天機器人於心理健康之倫理考量

Ethical Considerations for the Application of Chatbots in Mental Health

本文探討以大型語言模型為基礎之聊天機器人，應用於心理健康領域時所涉及之倫理議題。首先回顧現階段聊天機器人於心理健康領域的可能應用，繼而從醫學倫理四原則，不傷害、行善、尊重自主與正義，加上可解釋性（explainability）原則分析其在心理健康照護情境中可能引發之倫理挑戰，並進一步討論人工智慧介入對使用者所帶來的情感依附與情緒調節等心理影響，以反思此類技術之潛在益處、限制與風險。

This article examines the ethical issues arising from the use of large language model–based chatbots in mental health. It first reviews the current and potential applications of chatbots in mental health settings, and then analyses the ethical challenges they may generate through the lens of the four principles of biomedical ethics—non-maleficence, beneficence, respect for autonomy, and justice—supplemented by the principle of explainability. The discussion further considers the psychological effects of artificial intelligence–mediated interventions on users, including emotional attachment, emotional regulation, and transformations in the therapeutic relationship, to reflect critically on the potential benefits, limitations, and risks of these technologies.