When the artificial intelligence MOSS in the movie "Wandering Earth 2" identified hidden secrets by analyzing astronaut Liu Peiqiang's trembling lips and evasive gaze, this sci-fi scene not only shocked the audience, but also quietly revealed the possibility of emotional computing technology moving from imagination to reality. Nowadays, with the iteration of artificial intelligence technology, emotional computing is transforming from a screen plot to a key force in solving the dilemma of real-life psychological services, opening up a new track for the artificial intelligence industry that combines social value and market potential. Demand driven emotional computing breaks through the dilemma of mental health services. Emotional computing is a cutting-edge field that integrates computer science, psychology, sociology, and cognitive science. Its core is to empower computers with the ability to observe, understand, and generate human emotional expressions. Research shows that after more than 20 years of development, sentiment computing technology has evolved from early text sentiment analysis to multimodal analysis that integrates text, audio, and visual input, which can more comprehensively capture human emotional states. This technological advancement is precisely in line with the rapidly increasing demand for mental health services in today's society. Mental health issues are increasingly becoming a focus of attention globally. With the development of the economy and society and the intensification of competition, people are facing increasing pressure, and mental health issues are becoming more prominent. Especially in the area of adolescent mental health, which has received much attention, the current supply of psychological counseling services on campus is not yet mature and perfect, and there is still a contradiction of insufficient structural supply. In this context, "emotional computing empowered by artificial intelligence will become an emerging track," said Huang Minlie, a professor in charge of the Department of Computer Science and Technology at Tsinghua University, in an interview with reporters. "Artificial intelligence can not only think like humans, but also understand our emotions and feelings." Huang Minlie explained that on the one hand, many patients have a strong sense of shame towards mental illness, and the proportion of direct medical treatment is still very low; On the other hand, the consultation time is short and the amount of information does not meet the needs, making it difficult for psychologists to deeply understand the patient's needs. The continuous iteration of artificial intelligence technology will transform the form, efficiency, and effectiveness of psychological counseling and psychiatric diagnosis and treatment. Emotional computing, with its richness in multimodal data acquisition, convenience and accuracy in recognition, as well as advantages such as real-time adjustment, low threshold accessibility, and strong replicability, is becoming an important breakthrough in filling the gap in psychological services. Since the beginning of this year, we have been accelerating research and practice in emotional intelligence. Based on the rapid development of deep learning, generative artificial intelligence, and AI agents, artificial intelligence is expected to better understand human emotions, thinking, behavior, and cognition. ”Huang Minlie said. How can psychology and artificial intelligence algorithms "dance together" from "triangular modeling" to global industrial layout in technology implementation? Huang Minlie introduced that to solve a psychological problem with artificial intelligence, the first step is to model around three key psychological variables: human emotions, behavior, and cognition. Emotion refers to the ability to recognize, understand, and regulate one's own and others' emotions in social situations; Cognition is the fundamental ability to acquire, process, and apply information in response to others and social situations; Behavior is the fundamental ability to transform emotional and cognitive information into effective social actions and interactions. We found that these three key variables are interdependent triangles. Based on this, the second step is to use massive data to train an artificial intelligence model with cognitive modeling capabilities; The third step is to apply the trained large model to practical situations, based on understanding human cognition and thinking, to identify and interpret human behavior and emotions. This' modeling training application 'path provides a clear framework for the commercialization of sentiment computing. The reporter learned that many research institutions and technology companies in developed countries are currently accelerating their research in the field of emotional computing to explore the possibilities of applying artificial intelligence technology in emotional, psychological, and other fields. For example, Google's LaMDA, startup company Reflection AI's Pi, and startup company Character AI's Charact.AI are all based on their respective model technologies and research and development advantages, entering the emotional computing and emotional intelligence track. Domestic enterprises are also accelerating industrialization. For example, based on the domestically produced GLM-4.5 model, Lingxin Intelligence has launched Emoha, the first psychological model in China with empathy and chat capabilities. It has also launched three types of products, including an artificial intelligence psychological assessment system, AI dual teacher interactive psychological courses, and Lingxin Education Space, covering the complete service chain of "assessment intervention". When artificial intelligence maintains the level of a human consultant with 300 hours of experience, emotional intelligence has the preliminary ability to fill the gap in psychological service talent, "said Zheng Shuliang, CEO of Lingxin Intelligence. This is precisely the core goal of technology implementation: to use technological innovation to promote the popularization of mental health education. The inclusive prospects and ethical boundaries of emotional computing coexist with opportunities and challenges. The National Health Commission has released eight practical measures for serving the people's health by 2025, which stipulate that at least one hospital in each city should provide psychological and sleep clinic services; Promote the application of the national unified psychological assistance hotline 12356; More than 5000 national and provincial expert lectures on mental health knowledge have been held nationwide. Data shows that there are currently around 3000 comprehensive hospitals nationwide that can provide diagnosis, treatment, and rehabilitation services for mental health, but there are still many counties and districts with gaps in mental health services. The development and progress of artificial intelligence models in emotional intelligence are expected to promote equal rights in education and mental health services. ”Huang Minlie told reporters that with the current model capabilities, AI can provide accurate psychological state assessment and screening. As an AI consultant, AI can provide services, engage in AI training and supervision, and become a work assistant for psychological counselors. The rapid development of technology is also accompanied by ethical challenges. Industry experts say that emotional computing, which signals and models human emotions, may pose risks such as privacy breaches, algorithmic discrimination, and AI illusions. Experts believe that it is necessary to establish a dynamic consent model and a hierarchical regulatory path to find a balance between technological innovation and rights protection, such as how to protect emotional data privacy and avoid the harm of algorithmic bias to vulnerable groups. These are still the thresholds that the industry must cross. Looking ahead, emotional computing will present three major trends: deeper integration of multiple modalities to enhance the accuracy of emotion recognition; Enhance personalized service capabilities and provide customized support based on user background and personality traits; The human-machine collaboration model has matured, and AI has become an assistant to human consultants rather than a substitute, jointly building an efficient service system. From the sci-fi imagination of MOSS to the practical application of Emoha, emotional computing is rewriting the development trajectory of artificial intelligence. Its ultimate goal is not to replace human emotional interaction with machines, but to empower technology to make mental health services more accessible, efficient, and accurate, ultimately helping people's all-round development. On this new track, Chinese technology enterprises and research institutions have demonstrated strong competitiveness. With increased policy support, accelerated technological breakthroughs, and expanded application scenarios, emotional computing is expected to become a key area for the development of the artificial intelligence industry, contributing Chinese wisdom to the construction of the global mental health service system. (New Society)