Artificial intelligence (AI) is increasingly prevalent in mental health services, enhancing accessibility by providing immediate support through chatbots and remote platforms, and improving efficiency through automated diagnostics and personalized treatment recommendations. However, this rapid integration also brings numerous ethical controversies, including concerns over data privacy, algorithmic bias, and the potential erosion of human empathy in therapeutic relationships. This paper focuses on seven core ethical issues in AI mental interventions, including privacy protection, informed transparency, fairness and bias, responsibility attribution, autonomy and agency, emotional dependency, and simulated empathy. Existing studies mostly address single dimensions and fail to respond to the multi-stakeholder collaborative ethical challenges. To address this, the paper proposes a Layered Responsibility Framework that systematically analyzes the division of responsibilities and ethical constraints across three levels: developers, platform operators, and users. The study highlights that only by promoting clear accountability, transparent design, and institutional coordination can society ensure the sustainable application of AI technology in mental health and safeguard users psychological safety.
Xun Lei (Wed,) studied this question.