In the "AI Listening Room" at Central University of Finance and Economics, student Li Fei (pseudonym) shared his thoughts with an AI psychological counselor. She felt that in this soundproof, independent and soft light space, she was "unconditionally accepted". She did not have to worry about being judged or misunderstood, but could talk, roast, silence and cry. Xu Min, a full-time teacher at the Psychological Counseling Center of the Central University of Finance and Economics, introduced that "AI Lingxin Hut" is a university mental health management system based on large models and physical interaction spaces. It is a product of the center's deep cooperation with the Interactive Artificial Intelligence Research Group and its incubation enterprises at Tsinghua University, and will be officially launched in November 2025. In mid April, the team upgraded the "cabin" to improve AI response speed, opened the "mobile cabin" online and offline synchronous consultation mode, and added weekend consultation sessions. In April 2025, the Ministry of Education and nine other departments issued the "Opinions on Accelerating the Digitalization of Education", requiring the promotion of vertical application of large models in fields such as mental health. In February 2026, the Ministry of Education proposed the concept of "innovative digital intelligence empowerment" during the first Spring Festival meeting, deploying efforts to promote students' physical and mental health. In recent years, AI mental health services have entered many universities across the country. Yunnan University's "Smart Mental Health Home" attempts to provide online and offline spiritual care for teachers and students through digital technology; The "Heart Xiao" intelligent agent mini program independently developed by Shandong University of Science and Technology relies on the school's four level mental health work system to provide services... Campus AI mental health services are receiving increasing attention, with the advantages and risks being the focus issues. Feng Yi, Deputy Director and Associate Researcher of the Psychological Counseling Center at Central University of Finance and Economics, led a team to collaborate with Professor Qiao Zhihong's team from Beijing Normal University on conducting psychological research through the "AI Listening Room". They found that 92.86% of participating students believe that AI provides useful advice. The current version of the big model is equivalent to the level of a novice counselor with 50 hours of counseling, "said Fengyi. This means that the current big model can effectively solve some emotional problems, but cannot handle complex psychological counseling work. When it comes to the advantages of AI mental health services, Xiao Xiyao, the technical leader of the "AI Lingxin Hut" project, said that AI psychological counseling is based on the operation of multi-agent large models, with standardized information processing, and the convenience of AI mental health services is relatively high. However, its limitations are also evident, as AI still lacks in collecting multimodal information from counselors. For example, AI may have difficulty recognizing that counselors are silently crying, and AI psychological counseling cannot achieve a sense of "real person communication". Xu Min also mentioned that AI excels in structured methods such as cognitive-behavioral therapy, which can help students identify emotions and adjust negative thinking. However, AI still struggles to reach deeper levels of personality growth and self-awareness integration. She gave an example: "Mature consultants can identify deeper reasons from current problems, such as relationships with parents, which AI currently does not actively explore. ”Zhang Jing, a teacher at Communication University of China who has been engaged in psychological counseling and supervision for a long time, found in her research that the quality of AI mental health services varies greatly, and some AI models may use inappropriate language, which can backfire. She believes that attention should be paid to the privacy protection of AI mental health services and the establishment of risk warning mechanisms. The Psychological Counseling Center of Central University of Finance and Economics has conducted a layered design for risk warning. The first layer is during the system appointment stage, where risk screening is conducted through the school's mental health information database and questionnaire. Consultants with complex situations will be actively referred to real person counseling by the AI system. The second layer is that during the consultation process, AI integrates the evaluation process into natural dialogue. Once high-risk risks are identified, the system will pop up crisis intervention prompts and issue warnings to teachers through SMS or phone calls. The third layer is post consultation check, and teachers at the psychological counseling center need to regularly review the backend data of AI psychological counseling to confirm the user's crisis level. In Zhang Jing's view, policy empowerment and technological innovation provide space for the development of AI mental health services. In the future, it is necessary to build a human-machine collaborative mental health service system based on the principle of "assistance rather than substitution", seek a balance between technological innovation and ethical norms, and ultimately achieve the universality and precision of mental health services. Fengyi believes that with the continuous upgrading of AI technology, more universities will join the exploration and innovation of "AI+psychology", making it a new force in university mental health services. (Looking into the New Era)