EmoCare: Your AI diary that talks back with kindness
Since the emergence of generative AI several years ago, an increasing number of users have become deeply engaged in chatting with chatbots. However, studies indicate that these chatbots can sometimes offer biased, erratic, or even harmful advice to those seeking help. Furthermore, the echo chamber effect of chatbots may amplify users’ negative emotions and reinforce unfounded beliefs.
To address this issue, a team led by Professor Song Yanjie in the Department of Mathematics and Information Technology (MIT) developed an AI-powered application called EmoCare. This app allows users write diaries, generate pictures, chat with the AI companion to engage in self-reflection, and cultivate positive psychological states. A press conference was held in January 2026 to promote EmoCare and another app developed by the University.
“It is not uncommon for youth to be puzzled by emotional problems related to relationships, academic learning, and personal growth. EmoCare is designed to provide a platform for users to express their feelings through diary writing. The app then utilises Large Language Models (LLMs) and Natural Language Processing (NLP) to analyse diary entries and generate matching images and caring conversations,” explained Professor Song.
Developed by positive psychology pioneer Martin Seligman, PERMA encompasses five dimensions that contribute to well-being and happiness: Positive Emotion, Engagement, Relationships, Meaning, and Accomplishment. Drawing on the positive psychology framework of PERMA, EmoCare fosters a supportive environment for advice-seekers, encouraging them to express their thoughts and feelings while avoiding negative emotional spirals.
“When users document negative experiences, the app integrates refined positive elements before generating images and conversations to facilitate emotional regulation and prevent excessive immersion in negative emotional states,” said Professor Song, who also serves as the Associate Co-Director of Academy of Applied Policy Studies and Education Futures (AAPSEF). To enhance engagement, users can modify the images generated by EmoCare by providing additional input aligned with their emotions.
About 700 primary, secondary, and university students, including 68 students with special educational needs (SEN), have tried the EmoCare app. Findings revealed that users’ average happiness index increased by 1.881 points after using the app, which is 14.6% higher than that of an alternative app without the PERMA model. Among SEN students, nearly 94% showed improved emotional states after one month of continuous use.
Analysis also revealed that SEN students particularly valued recognition in the dimension of “Accomplishment.” Professor Song explained, “Based on these findings, we plan to strengthen support for SEN students in their pursuit of accomplishment and design interaction models that better align with their learning and emotional development.”
The data visualisation platform helps professionals closely monitor the emotional changes of individuals under their care.
The application is equipped with an emotional analytics function to collect and analyse users’ emotional data and track trends. It also features a built-in dashboard that provides a visual platform displaying the emotional analytics. “The data visualisation platform helps educational psychologists and counselling professionals closely monitor the emotional changes of individuals under their care and identify those at risk,” Professor Song added.
Mr William Chow, Psychological Counsellor at Hong Kong Psychological Counselling Center, commented, “EmoCare has been developed in line with the interests and needs of youth in their mental growth. Users can express the joys and sorrows of daily life through writing, while the platform, grounded in the PERMA model, responds with AI‑generated images. This form of interaction carries elements of psychological counselling, and with timely intervention from counsellors, the effectiveness of guidance can be significantly enhanced.”
Mr Kam Wai Ming, Principal of HKSKH Bishop Hall Secondary School, who adopted EmoCare for about 120 students, added, “While AI certainly enhances students’ learning ability, the distinctive value of EmoCare lies in its simultaneous focus on students’ emotions and growth, making education warmer and more human-centred”.
Professor Sin Kuen Fung, Executive Director of the Institute of Special Needs and Inclusive Education at EdUHK, further noted, “Physical and mental well-being begins with awareness. EmoCare helps young people with special educational needs learn to observe, reflect, and respond with care. With appropriate support and guidance in self-regulation, they can develop resilience and enjoy broader inclusion in social circles”.
EmoCare was selected for the EdUHK-HKSTP “Co-Ideation Programme” in October 2025. The knowledge-transfer project has already been showcased at the Belt and Road International Conference on Special Education 2025 and at InnoCarnival 2025. With features such as portable printers that instantly produce diary content and accompanying images, the app has been well received by students, parents, and educators alike. At the same time, MetaAI, a start-up founded by Professor Song, is actively seeking industry partners to explore commercialisation applications, aiming to bring its innovations to a broader community of students in Hong Kong, the Greater Bay Area, and beyond.
Standing at the forefront of innovation, we will remain committed to our research and knowledge-transfer work.
Professor Song said she owes much credit to her research team who have been working closely together on a number of innovative projects aimed at enhancing educational experiences through technology. “Together, we have invented several award-winning applications, including the commercialised metaverse platform Learningverse, the AR and VR-supported the vocabulary learning app VocabGo, the AI-powered EmoCare app, and newly developed iChat system featuring customised digital humans for learning and teaching purposes, and Embodyverse, which integrates an embodied AI-driven robot in mixed reality (MR) for collaborative science learning. Standing at the forefront of innovation, we will remain committed to our research and knowledge-transfer work,” Professor Song said.
Click here to review a news release about other AI-related research projects conducted by Professor Song’s team.



