Nutzung künstlicher Intelligenz zur Optimierung psychologischer Traumaintervention in der Sozialhilfe

TYPE Original Research PUBLISHED 21 April 2026 DOI 10.3389/fpsyg.2026.1578545 Utilizing artificial intelligence to OPEN ACCESS optimize psychological trauma EDITED BY Pietro Cipresso, intervention in social assistances University of Turin, Italy REVIEWED BY Yu Zhang 1†, Wenzuixiong Xiong 2*† and Yuxiang Gao 2† Antonia Cascales-Martinez, University of Murcia, Spain 1 School of Law and Economics, Wuhan University of Science and Technology, Wuhan, China, 2School Luis Fabián Salazar-Garcés, of Art, Hubei University, Wuhan, China Technical University of Ambato, Ecuador *CORRESPONDENCE Wenzuixiong Xiong 20220132@hubu.edu.cn Introduction: Psychological trauma is prevalent among beneficiaries of social † These authors have contributed equally assistance, posing significant challenges to effective intervention. Traditional psy- to this work and share first authorship chological support methods are often constrained by limited resources, delayed RECEIVED 17 February 2025 responses, and insufficient personalization. Against this backdrop, artificial intel- REVISED 21 February 2026 ligence (AI) offers new opportunities to enhance the efficiency and precision of ACCEPTED 31 March 2026 PUBLISHED 21 April 2026 psychological trauma interventions. This study explores the feasibility and effec- tiveness of integrating AI technologies into social assistance systems. CITATION Zhang Y, Xiong W and Gao Y (2026) Methods: This study adopts a mixed analytical approach. First, it reviews the cur- Utilizing artificial intelligence to optimize rent applications of AI in mental health, focusing on natural language process- psychological trauma intervention in ing, emotion recognition, and personalized recommendation systems. Based on social assistances. Front. Psychol. 17:1578545. this, an AI-driven framework for psychological trauma intervention is proposed, doi: 10.3389/fpsyg.2026.1578545 including the development of intelligent auxiliary diagnostic systems and per- COPYRIGHT sonalized intervention plans. Empirical data are collected through questionnaire © 2026 Zhang, Xiong and Gao. This is surveys to evaluate user experiences and intervention outcomes. an open-access article distributed under Results: The findings indicate that AI-assisted interventions significantly improve user the terms of the Creative Commons Attribution License (CC BY). The use, satisfaction and mental health outcomes. Specifically, 75% of respondents report being distribution or reproduction in other satisfied or very satisfied with the intelligent mental health system. Approximately forums is permitted, provided the 60% of participants experience noticeable improvements in their mental health status. original author(s) and the copyright owner(s) are credited and that the Furthermore, 85% of respondents consider the personalized intervention plans to be original publication in this journal is well-aligned with their individual needs and circumstances. cited, in accordance with accepted Discussion: The results demonstrate that AI technologies can effectively address academic practice. No use, distribution or reproduction is permitted which does key limitations of traditional psychological interventions by enhancing accessibil- not comply with these terms. ity, responsiveness, and personalization. The integration of AI into social assis- tance systems not only improves intervention outcomes but also offers scalable solutions for resource-constrained settings. This study provides empirical sup- port for the application of AI in social assistance and highlights its transformative potential in advancing mental health governance. KEYWORDS artificial intelligence, intelligent auxiliary, diagnostic system, privacy protection, psychological trauma, social assistance 1 Introduction In the context of the new era, social assistance is an important social activity. It aims at providing necessary support and resources to vulnerable groups to improve their quality of life and social integration (Hou et al., 2020; Mishra, 2020; Shao et al., 2020). However, in this process, people often overlook an important issue: beneficiaries may face various psychological traumas, posing chal- lenges to assistance work. Psychological traumas, such as post-traumatic stress disorder (PTSD), anxiety, and depression. They affect the mental health of beneficiaries and may hinder their social Frontiers in Psychology 01 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 integration and the process of rebuilding their lives (Jolly et al., 2020). influence their emotions and behaviors. By identifying and challeng- Traditional psychological intervention methods, such as counseling and ing negative thought patterns, individuals can learn more positive and therapy, although can alleviate psychological traumas to some extent, are adaptive coping strategies. Affective computing is a field of study plagued by many issues. Resource scarcity, insufficient services, and lack focused on how to enable computers to understand and generate emo- of professional personnel manpower result in many beneficiaries not tions. By analyzing users’ behavior, speech, and physiological signals, receiving timely and effective help. Especially for vulnerable groups in affective computing can recognize and understand users’ emotional social assistance, due to the uniqueness of their experiences and the diver- states, providing more personalized support and services. AI systems sity of their needs, traditional psychological intervention methods often can leverage technologies such as speech recognition, facial expression prove inadequate (Alam, 2022; Fu et al., 2020; Zhang et al., 2020). recognition, and physiological signal monitoring to accurately identify With the continuous advancement of technology, artificial intelli- changes in users’ emotions. These technologies can help the system gence (AI) technology has gradually become a new choice for optimizing promptly detect emotional issues and offer appropriate support. As AI psychological trauma intervention in social assistance (Sun et al., 2020). technology continues to advance and develop, more and more studies Natural language processing (NLP) technology can analyze the linguistic are beginning to explore the possibility and effectiveness of utilizing expressions of beneficiaries. Emotion recognition technology can identify AI technology to optimize psychological trauma intervention in social changes in beneficiaries’ emotions. Personalized recommendation tech- assistance. For instance, Parker and Grote (2019) designed interven- nology can recommend corresponding intervention plans based on the tion programs targeting different types of psychological trauma using characteristics of beneficiaries. The application of these technologies pro- personalized recommendation technology, achieving good results. vides new avenues and possibilities for improving mental health (Torous These studies suggest that AI technology provides new ideas and et al., 2021; Xu et al., 2021). When services are provided solely by humans, methods for optimizing psychological trauma intervention. Leung et issues such as uneven resource distribution, shortage of professionals, and al. (2021) explored the application of AI-based virtual psychotherapy slow response time may arise. This is especially problematic in remote in their research. They developed a virtual psychotherapy system that areas or during emergencies, where mental health services often struggle analyzed users’ speech and emotions through NLP techniques, and to meet demand in a timely and effective manner. provided personalized psychological support and advice based on the This work aims to explore the feasibility and effectiveness of utilizing analysis results. The research results indicate that the system effectively AI technology to optimize psychological trauma intervention in social alleviates users’ anxiety and depression symptoms, improving their assistance. First, this work analyzes the current situation of psychological mental health levels. This study provides new ideas and methods for trauma in social assistance and discusses the limitations of traditional using AI technology in psychotherapy. psychological intervention methods. Next, this work introduces the cur- Mylona et al. (2022) primarily focused on the analysis of psycho- rent application status and potential advantages of AI technology in the logical counseling dialogues based on NLP technology. They analyzed field of mental health. Then, it proposes methods and strategies to opti- the speech expressions of beneficiaries and counselors in psychologi- mize psychological trauma intervention in social assistance using AI tech- cal counseling dialogues and studied the interaction patterns and nology. Finally, this work looks forward to the future prospects of AI in influencing factors between the two. The study found that the linguis- the field of social assistance and emphasizes the importance of addressing tic features in psychological counseling dialogues were closely related ethical and privacy protection issues while advancing technology. This to the psychological state of the beneficiaries and could be used to work is expected to provide new ideas and methods to improve the predict and assess their mental health levels. This study provides coun- mental health of beneficiaries and contribute to the further development selors with better tools to understand the needs of beneficiaries, thus and improvement of social assistance work. helping to improve the effectiveness and quality of psychological counselling. Barbalat et al. (2024) concentrated on designing psycho- logical intervention plans using personalized recommendation tech- nology. They developed a personalized intervention system based on 2 Literature review machine learning algorithms, recommending corresponding psycho- logical intervention plans based on the characteristics and needs of the In the field of mental health, the application of AI technology has beneficiaries. The research results showed that personalized interven- become increasingly widespread. NLP techniques can analyze the speech tion plans are more effective in alleviating psychological trauma and of beneficiaries, emotion recognition technology can identify changes in improving mental health compared to traditional generic intervention the emotions of beneficiaries, and personalized recommendation technol- plans. This study provides new ideas and methods for improving the ogy can recommend appropriate psychological intervention programs specificity and effectiveness of psychological intervention. Fei et al. based on the characteristics of the beneficiaries. The application of these (2020) aimed to monitor mental health conditions using emotion rec- technologies provides new avenues and possibilities for improving mental ognition technology. They developed an emotion recognition system health (Rowe and Lester, 2020). For example, Zhang et al. (2022) utilized that identified users’ emotional states by analyzing their speech and emotion recognition technology to analyze the speech and behavior of behavior and provided corresponding psychological support and social assistance recipients, timely identifying and intervening in the advice based on emotional changes. The research results demonstrated development trends of psychological problems, achieving certain effec- that the system accurately identified users’ emotional changes and tiveness. Additionally, Goldberg et al. (2020) studied an NLP-based effectively improved their mental health levels. This study provides method for analyzing psychological counseling dialogues, providing new avenues and possibilities for using emotion recognition technol- counselors with better tools to understand the needs of beneficiaries. ogy for mental health monitoring and intervention. Emotional Cognitive Behavioral Therapy (CBT) is a widely used psychologi- changes are a crucial basis for AI systems to provide support and rec- cal treatment method aimed at improving mental health by changing ommendations. The system uses NLP and emotion recognition tech- an individual’s thought patterns and behavioral habits. The core nologies to conduct a comprehensive analysis of the beneficiary’s assumption of CBT is that an individual’s thoughts and beliefs directly voice, facial expressions, and physiological indicators to accurately Frontiers in Psychology 02 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 identify their emotional state and mental health condition. Specifically, and on the potential problems in the service process. This can ensure the system looks for emotional vocabulary, tone variations in the that clients receive comprehensive and effective psychological voice, microexpressions, and muscle activity in the face, and monitors support. physiological indicators such as heart rate and blood pressure to com- prehensively assess emotional changes. These emotional changes are (1) Concept and Characteristics of Psychological Trauma in Social calculated using algorithmic models, which quantify the emotional Assistance: Psychological trauma in social assistance refers to state based on the beneficiary’s historical and real-time data, providing the phenomenon of psychological trauma occurring during a foundation for subsequent support and recommendations. the process of receiving social assistance. This trauma is usu- The research conducted by these scholars provides crucial theo- ally caused by individuals experiencing a series of negative retical and practical support for the application of AI technology in events or hardships, such as domestic violence, natural disas- the field of mental health. Their work enables a better understanding ters, and unemployment, leading to psychological distress. and utilization of AI technology, providing new ideas and methods for Social assistance recipients often find themselves in disadvan- improving mental health. In the future, people can further draw on taged positions, lacking social support and resources, making their research findings to continuously improve related technologies them more susceptible to the influence of external events and and methods, making greater contributions to the development of the thus experiencing psychological trauma. field of mental health. Moreover, with the continuous progress and (2) Types and Manifestations of Psychological Trauma: In the application of AI technology, there is the opportunity to delve deeper context of social assistance, psychological trauma can manifest into research, explore further, continuously refine relevant technolo- in various types, including PTSD, anxiety, and depression. gies and methods, and make greater contributions to enhancing the PTSD is a common type of psychological trauma, character- mental health of beneficiaries. However, the process of utilizing AI ized by persistent negative emotions and behavioral reactions technology for psychological intervention also needs to pay attention triggered by traumatic events. Anxiety and depression are also to ethical and privacy protection issues to ensure the benign applica- common emotional issues, characterized by sustained feelings tion of technology, and truly benefit the recipients of social assistance. of tension, fear, and despondency. These types of psychologi- A review of relevant literature demonstrates that AI technology cal trauma are prevalent among social assistance recipients, provides new ideas and methods for psychological trauma interven- severely impacting their mental health and quality of life. tion in social assistance, carrying significant theoretical and practical (3) Mechanisms and Influencing Factors of Psychological Trauma: implications. The occurrence mechanism of psychological trauma in social assistance is a complex process influenced by individual, envi- ronmental, and societal factors. Individual factors include mental health status and coping strategies; environmental fac- 3 Methods tors encompass family environment and social support; and societal factors involve socioeconomic status and cultural 3.1 Theoretical analysis of psychological background. These factors interact with each other, collec- trauma in social assistance tively affecting the individual’s risk and severity of psychologi- cal trauma. Psychological trauma is often closely related to the Social assistance aims to provide necessary support and events individuals experience, such as accidents, natural disas- resources to individuals in need, to improve their quality of life and ters, wars, and violent incidents. The suddenness, severity, and social integration. However, individuals receiving social assistance unpredictability of these events can cause immense psycho- often experience various misfortunes and pressures, which may lead logical shock to individuals, leading to trauma. Factors such to the emergence of psychological trauma. Theoretical analysis of as an individual’s psychological resilience, coping mecha- psychological trauma in social assistance helps to deepen under- nisms, and personality traits play significant roles in the standing of its mechanisms and influencing factors, providing a occurrence of psychological trauma (Mitchell, 2019). The theoretical basis and guidance for psychological intervention (Boyer, more fragile, vulnerable, or lacking effective coping strategies 2022; Qiu and Wu, 2024; Saiyu et al., 2022; Mitchell, 2019). When an individual is, the more susceptible they are to the effects of discussing the concept and characteristics of psychological trauma psychological trauma. The adequacy of social support systems in social assistance centers, it is important to clearly distinguish plays a crucial role in the occurrence and impact of psycho- between the traumatic events themselves and the process of receiv- logical trauma. Lack of social support, adverse environments, ing social assistance. In fact, most clients, after experiencing trau- social conflicts, and other factors exacerbate the severity of matic events (such as natural disasters, violence, abuse, and loss of individual psychological trauma. An individual’s cognition loved ones), actively or passively seek social assistance services to and interpretation of events also affect the occurrence of psy- obtain necessary support, aid, and recovery. These traumatic events chological trauma. For example, different interpretations of are the direct causes of the client’s psychological trauma, not the the same event may result in varying degrees of trauma, and process of receiving social assistance itself. In some cases, clients may positive cognitive approaches can alleviate the severity of encounter issues such as communication barriers, inadequate ser- trauma. Psychological trauma is not merely a psychological vices, and lack of trust while receiving social assistance. While these state but also affects the individual’s physiological functions. issues are not directly equivalent to the traumatic events themselves, Prolonged psychological trauma may lead to a series of physi- they can negatively impact the client’s psychological state and exac- cal problems, such as insomnia, headaches, and digestive erbate their trauma. Therefore, when providing social assistance issues, thereby affecting the individual’s quality of life (Zhang services, it is essential to focus on the traumatic events themselves et al., 2020). Frontiers in Psychology 03 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 3.2 The use of AI in the field of mental convert voice input into text. It also adopts OpenCV along with deep health learning models to recognize users’ facial expressions. (4) Emotion Recognition Module: By integrating the results of NLP, speech rec- Traditional psychological intervention methods, such as counsel- ognition, and facial expression analysis, the system comprehensively ing and therapy, although to some extent alleviate psychological evaluates users’ emotional states, monitors real-time changes in their trauma, are fraught with various issues. For instance, resource scarcity, emotions, and records historical emotional data. (5) Personalized insufficient services, and a shortage of professional manpower often Recommendation Module: Based on users’ personal information, result in many recipients not receiving timely and effective assistance mental health status, and historical data, the system constructs users’ (Sanderson et al., 2020; Sloshower et al., 2019). However, with the profiles and recommends relevant mental health articles, videos, and rapid development and continuous innovation of AI technology, its exercises according to users’ interests and needs. (6) Data Storage application in the field of mental health is becoming increasingly and Management Module: The system encrypts user data using algo- widespread. It offers new avenues and methods for the diagnosis, rithms like AES to ensure data security. (7) Assessment and intervention, and treatment of mental health issues. AI technology, Feedback Module: The system periodically sends satisfaction surveys with its powerful data processing and analysis capabilities, has shown to users, collects feedback, evaluates users’ mental health using tremendous potential in the field of mental health, providing psy- mental health scales, and tracks changes before and after interven- chologists, doctors, and patients with more intelligent and personal- tions. Figure 2 illustrates the application of AI in the field of ized services (Alowais et al., 2023; Doraiswamy et al., 2020; Lee and mental health. Yoon, 2021). The Intelligent Mental Health Assistance System (IMHAS) is an (1) NLP Technology: The application of NLP technology in the integrated platform that combines NLP, emotion recognition, per- field of mental health is becoming increasingly widespread. By sonalized recommendation, and data mining technologies. The flow analyzing the speech expressions of beneficiaries, NLP tech- chart of IMHAS algorithm is shown in Figure 1. The system is nology assists psychologists and doctors in better understand- designed to provide personalized mental health support and inter- ing patients’ inner worlds and emotional states, thus providing vention plans for individuals in need of social assistance. IMHAS more personalized and effective psychological support and consists of the following main modules. (1) User Interface Module: intervention. For example, by analyzing the content and emo- Users can access the system through various channels such as web tional tone of patients’ speech, NLP technology can automati- portals, mobile apps, and WeChat mini-programs. The system cally generate psychological assessment reports and records users’ personal information and usage history. (2) Data intervention plans, providing scientific evidence for the diag- Collection Module: The system collects data from users via text nosis and treatment of mental health issues (Le Glaz et al., input boxes, voice input, and camera-captured facial expressions. It 2021). This work employs deep learning-based NLP models, also gathers physiological signal data through wearable devices. (3) such as BERT and Transformer, to perform sentiment analysis Data Processing Module: The system employs deep learning models and semantic understanding of users’ text input. These models like BERT and Transformer for sentiment analysis, semantic under- can identify users’ emotional states, extract keywords, and standing, and keyword extraction of users’ text input. It uses the detect underlying themes, enabling more accurate psychologi- Google Speech-to-Text API or other speech recognition engines to cal assessments. FIGURE 1 IMHAS algorithm flow chart. Frontiers in Psychology 04 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 FIGURE 2 The application of AI in the field of mental health. (2) Emotion Recognition Technology: It is another crucial appli- prioritize disease prevention and rehabilitation care. Gender cation of AI technology in the field of mental health. By ana- differences are also significant in health needs. Women may lyzing patients’ speech, facial expressions, and physiological be more concerned with breast health, prenatal care, and indicators, emotion recognition technology can accurately menopause management, whereas men may focus more on identify patients’ emotional states and mental health condi- prostate health, cardiovascular health, and similar areas. The tions, providing timely and effective support for psychological machine algorithm provides targeted health recommenda- intervention and treatment (Gual-Montolio et al., 2022). For tions based on these gender differences. An individual’s life- example, emotion recognition technology can monitor style habits play a crucial role in their health. Unhealthy changes in patients’ emotions, and promptly identify and behaviors, such as smoking, alcohol consumption, and staying intervene in the development trends of mental health issues, up late, increase the risk of illness. The machine algorithm thereby improving the accuracy of prediction and diagnosis of offers health improvement suggestions based on these lifestyle mental health problems. This work utilizes speech recognition, habits and encourages the adoption of healthier living prac- facial expression recognition, and physiological signal moni- tices. The system employs collaborative filtering and deep toring (such as heart rate and skin conductivity) to compre- learning algorithms to recommend personalized psychological hensively assess changes in users’ emotions. These technologies intervention plans based on users’ personal characteristics and are trained using machine learning algorithms, enabling accu- historical data. The recommendation system also continuously rate identification of users’ emotional states across different optimizes its suggestions based on user feedback. contexts. (3) Intelligent-assisted Diagnosis System: The intelligent-assisted 3.3 Methods and strategies for optimizing diagnosis system is another important application of AI tech- psychological trauma intervention in social nology in the field of mental health. By utilizing machine assistance centers using AI learning and data mining techniques to analyze a large amount of mental health data, this system can assist doctors in diag- The effectiveness and quality of psychological trauma intervention nosing various mental health problems quickly and accurately, in social assistance centers are directly related to the mental health and and provide personalized treatment plans for patients (Chen social integration of vulnerable groups. Traditional psychological et al., 2021). For example, it can predict patients’ disease risks intervention methods have various limitations, while the application and treatment outcomes based on their symptoms and mental of AI technology provides new ideas and methods for optimizing psy- health history, providing scientific evidence for clinical chological trauma intervention in social assistance centers. This sec- decision-making. tion explores methods and strategies for optimizing psychological (4) Personalized Recommendation System: The personalized rec- trauma intervention in social assistance centers using AI technology. ommendation system is another crucial application of AI tech- In a certain social assistance institution, to better help beneficia- nology in the field of mental health. By analyzing patients’ ries cope with mental health issues, they have introduced an IMHAS personal characteristics and health needs, this system can intelligent mental health assistance system. This system, based on AI provide personalized mental health services and treatment technology, combines NLP, emotion recognition, and intelligent dia- plans for patients, thereby improving treatment effectiveness logue technologies to provide personalized psychological support and and patient satisfaction. Individuals of different age groups intervention for beneficiaries. The process of system evaluation exhibit variations in their physiological, psychological, and involves three main steps: data preprocessing, feature extraction, and social needs. Younger individuals may be more focused on model prediction. First, the system preprocesses the raw data from physical fitness and mental health, while older adults may beneficiaries, including voice, facial expressions, and physiological Frontiers in Psychology 05 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 indicators, through operations such as noise reduction, filtering, and wearable devices, providing real-time, personalized health suggestions normalization. These steps improve the quality and accuracy of the and interventions. The system also offers remote medical consulta- data. Next, the system extracts feature related to emotional changes, tions and expert advice services to help clients better understand and such as the frequency of emotional vocabulary, amplitude of tone implement the recommended interventions. When selecting interven- variation, and key points in facial expressions. These features are used tion measures, multiple factors are considered, including the effective- in subsequent model predictions. Finally, the extracted features are fed ness, safety, feasibility, and client acceptance of the measures. Through into a trained algorithm model. The model predicts and assesses the iterative refinement and optimization, the system gradually improves emotional state of the beneficiary based on these features, resulting in the accuracy and satisfaction of the intervention matches. a quantified evaluation of emotional changes. Figure 3 illustrates the methods and strategies for optimizing psychological trauma interven- (1) Personalized Intervention Plan Design: This system first con- tion using AI. ducts personalized assessments of beneficiaries, collecting When constructing the AI health management system, particular information such as their personal characteristics, mental attention is paid to matching individual preferences with appropriate health status, and preferences. Based on the beneficiary’s emo- intervention measures. Besides, corresponding algorithms are tional changes and mental health status, the AI system pro- designed to ensure that the system can accurately recommend inter- vides personalized support and recommendations. These ventions tailored to clients’ needs. First, a user profiling approach is supports and suggestions mainly include encouragement, adopted, wherein personal information, health data, and historical comfort, and psychological counseling, aimed at helping ben- behavior records are collected and analyzed to create a personalized eficiaries alleviate negative emotions such as anxiety and profile for each client. This profile includes various dimensions of depression, while improving their overall mental health. The information such as the client’s health status, disease risk, lifestyle system automatically selects the most suitable support and habits, and interests, providing a foundation for matching subsequent suggestions to deliver based on the beneficiary’s emotional intervention measures. Next, a comprehensive intervention measure state and needs. For example, when the system detects a sig- database is established, containing a wide range of interventions for nificant level of anxiety in the beneficiary, it will automatically different health issues and needs. These measures include, but are not generate and push encouraging and supportive messages. limited to, medication treatments, physical therapy, psychological Similarly, when the system identifies a need for psychological interventions, nutritional guidance, and exercise recommendations. counseling, it will recommend professional counselors or psy- This database is not a static, pre-loaded list but is continuously chological intervention resources for the beneficiary to updated and refined based on the latest medical research and practical choose from. experience. (2) Emotion Recognition and Mood Monitoring: The system ana- During the matching process, the system selects the interventions lyzes the speech expressions and emotional states of beneficia- that best align with the client’s needs and preferences from the inter- ries to monitor their emotional changes and mental health vention database, based on a multi-factor comprehensive evaluation. status in real-time. For example, when the system detects This matching process is based on a multi-factor comprehensive eval- negative emotional vocabulary or tone in the beneficiary’s uation, which includes, but is not limited to, the client’s health status, speech, it immediately alerts the psychological counselor for disease risk, lifestyle habits, interests, and specific health goals and intervention. Meanwhile, the system also records the history expectations. of the beneficiary’s emotional changes, providing more com- Finally, the AI system delivers these intervention measures to cli- prehensive and accurate information support for psychologi- ents through multiple channels, such as mobile apps and smart cal intervention. FIGURE 3 Methods and strategies for optimizing psychological trauma intervention using AI. Frontiers in Psychology 06 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 (3) Intelligent Dialogue System: Through intelligent dialogue generate emotion reports, providing healthcare professionals with technology, the system communicates and interacts with ben- timely emotional monitoring and intervention support. eficiaries, providing personalized psychological support and Conducting a questionnaire survey is a crucial means to evaluate advice. For example, the system can automatically generate the effectiveness and user satisfaction of the IMHAS in psychological corresponding intervention content based on the beneficiary’s trauma intervention in social assistance centers. When designing the speech expressions and emotional states, such as encourage- questionnaire survey, several dimensions need to be considered: ment, comfort, and advice, to help beneficiaries understand and cope with mental health issues. (1) Satisfaction Evaluation: it is to evaluate the overall satisfaction (4) Data-Driven Intervention Strategies: By analyzing the mental of the respondents with the IMHAS and the satisfaction with health data and behavioral data of social assistance recipients, various functions, including the usability, practicality, and the system extracts valuable information and patterns to pro- effectiveness of the system. Satisfaction primarily reflects the vide scientific evidence for psychological intervention. For beneficiaries’ overall evaluation and user experience with the example, the system can identify potential mental health issues system, while effectiveness assesses the system’s impact on and risk factors, and propose corresponding intervention rec- improving the beneficiaries’ mental health status. To measure ommendations and measures by analyzing the psychological client satisfaction with the AI system, a comprehensive cus- assessment data and social network behavior data of tomer satisfaction survey is designed. This questionnaire beneficiaries. covers various aspects such as the system’s ease of use, user interface friendliness, response time, and level of personaliza- The above examples demonstrate the application of AI technology tion. It aims to gain a thorough understanding of the client’s in psychological trauma intervention in social assistance centers. By feelings and experiences during system usage. When evaluat- introducing intelligent assistance systems, social assistance institutions ing the effectiveness of the intervention measures, corre- can provide beneficiaries with more personalized, timely, and effective sponding evaluation standards and indicators are established psychological support and intervention, improving the targeting, based on the client’s health status, disease type, and the specific effectiveness, and scientificity of psychological intervention. This content of the intervention. These standards and indicators system provides new ideas and methods for the improvement and may include improvements in physiological indicators (such enhancement of social assistance work. as blood pressure and blood glucose levels), improvements in psychological states (such as anxiety and depression), and 3.4 Questionnaire survey enhancements in quality of life. (2) Intervention Effect Assessment: This is to evaluate the effec- This work employs a variety of methods to comprehensively ana- tiveness of the IMHAS in intervening in psychological trauma. lyze the content and emotional tone of patent presentations, and This includes assessing whether the respondents feel an monitor changes in patients’ emotions. First, patent presentation improvement in their mental health status and whether they data relevant to this work are collected from multiple authoritative find it easier to cope with difficulties and challenges. patent databases. These datasets cover patent presentations from dif- (3) Personalization Degree Evaluation: This is to evaluate the sat- ferent fields and time periods, ensuring the diversity and represen- isfaction of the respondents with the personalized interven- tativeness of the sample. Strict privacy protection and data security tion plans provided by the system. This includes assessing protocols are followed during data collection to ensure the legality whether the system provides corresponding intervention plans and compliance of all data. To evaluate the intervention effect of based on individual characteristics and needs, and the speci- IMHAS on psychological trauma, an independent-samples t-test is ficity and practicality of the intervention plans. used to analyze the intergroup differences in Patient Health (4) Interaction Experience Evaluation: This is to evaluate the Questionnaire-9 (PHQ-9) and Generalized Anxiety Disorder-7 experience of the respondents with the system’s intelligent dia- (GAD-7) scores before and after the intervention to examine the logue and emotion recognition functions. This includes significance of differences in intervention effects between the two assessing whether the interaction with the system is smooth, groups. A paired-samples t-test is used to analyze the intragroup and whether the system can accurately recognize and under- differences in scores before and after the intervention in each group stand the users’ emotional states. to examine the significance of the intervention effect within a (5) Data Privacy Protection Assessment: This is to evaluate the single group. system’s data privacy protection measures. This includes To analyze the content and emotional tone of the patent presenta- assessing whether the system can protect users’ personal pri- tions, NLP techniques are employed. Initially, text preprocessing steps vacy information and whether there are sufficient data secu- such as tokenization and part-of-speech tagging are used to clean and rity measures. standardize the patent presentations. Then, sentiment analysis algo- rithms are applied to quantify the emotional tendencies within the The participants in this study must meet the following criteria. presentations, including positive, negative, and neutral emotions. They are aged between 18 and 65, voluntarily participate in the study, Additionally, methods like keyword extraction and topic modeling are are capable of using a smartphone or computer, and agree to sign an used to analyze the core content and key information of the patent informed consent form. Participants with severe cognitive impair- presentations. To monitor changes in patients’ emotions, an emotion ments, mental illnesses, or those unable to cooperate in completing recognition system based on multimodal data from patient voice, the study are excluded. All participants’ personal information is kept facial expressions, and text communication is designed. This system strictly confidential, and data are stored on encrypted servers, acces- can capture and analyze patients’ emotional changes in real-time and sible only by authorized researchers. The system employs Frontiers in Psychology 07 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 anonymization to ensure that users’ private information is not dis- time, and meanwhile improve their cognitive ability of technology closed. A high priority is placed on protecting participants’ data pri- through the AI psychological knowledge popularization module in vacy. Before recruiting participants, the purpose of the study, and the the system, thus realizing technological empowerment. methods of data collection and use, are explained to them in detail; Participants are recruited through various channels. First, recruit- written consent is also obtained to ensure their voluntary participation ment information is posted in communities, medical institutions, and after fully understanding the study’s content and risks. Two hundred online platforms, clearly outlining the study’s purpose, requirements, social assistance recipients participate in the questionnaire survey. participants’ rights, and protection measures. Next, collaborations are Two hundred questionnaires are distributed, 190 are collected, and established with local medical institutions to invite eligible patients to after excluding invalid questionnaires, 180 valid questionnaires are participate. Finally, recruitment invitations are sent to potential par- obtained, with a recovery rate of 94.7%. To more comprehensively ticipants via social media and email. To ensure the diversity and rep- validate the effectiveness and advantages of the IMHAS in psychologi- resentativeness of the participants, clear recruitment criteria and cal health interventions, this work designs a control group with a total screening processes are established. These criteria cover various of 190 participants. One group receives services provided by the AI aspects such as age, gender, and health status to ensure that the par- system, while the other group receives services from professional ticipants can accurately reflect the characteristics of the target popula- healthcare providers. The two groups are balanced in terms of age, tion. During data collection, identical evaluation standards and gender, and health status to ensure fairness and accuracy in the com- questionnaires are used to objectively assess the service effects for both parison results. groups. These evaluation standards cover improvements in health The social assistance recipients vary significantly in age, educa- indicators, service satisfaction, the effectiveness of intervention mea- tional level and digital literacy. If the model training data fails to fully sures, and other aspects. Comparing the data from both groups allows cover all groups, algorithmic bias may arise. This study optimizes the for a clearer understanding of the differences and advantages between sample structure through stratified sampling and adds a fairness- the AI system and human-provided services in health management. constrained loss function to model training to reduce the performance disparities among groups. Given that psychological interventions involve sensitive information such as users’ core emotional experi- ences and traumatic memories, any leakage of such information may 4 Results cause secondary harm. In addition to adopting AES encryption for data storage, this study further implements strategies including the 4.1 Basic information of survey participants principle of data minimization, hierarchical access rights and data lifecycle management. In view of the possibility that some assistance Before assessing the satisfaction and intervention effects of the recipients passively accept AI interventions due to factors such as the IMHAS, this work conducts statistical analysis on the gender, age digital divide and psychological dependence, this study safeguards group, and educational background of the survey participants. their right to independent choice through a dual-track intervention Figure 4 shows the basic information of the survey participants. model. Users can switch between AI and human interventions at any Among the participants, the proportion of females is larger, FIGURE 4 Basic information of survey participants. Frontiers in Psychology 08 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 accounting for 70%, while males make up 30%. This suggests that 4.3 Intervention effect evaluation women are more proactive or willing to participate in the survey regarding the use of the system. The majority of participants are aged This work assesses the intervention effects on the survey partici- between 18 and 30, making up 44.45%. This likely reflects a growing pants. After using the system, 60% of the respondents report an awareness of mental health issues among young people, who tend to improvement in their mental health condition. 25% of the respon- seek help through technological means. Participants aged 31 to 45 dents feel more confident and optimistic when facing difficulties and account for 33.33% of the total, indicating that individuals in this age challenges. Through comparative analysis, it is found that respon- group often face dual pressures from their careers and families. This dents who demonstrate more confidence and optimism when facing potentially leads to a greater demand for mental health services. challenges typically have simpler or more easily controllable symp- Participants aged 46 and above make up 22.22% of the total. It sug- toms. These individuals are more likely to accept and adapt to the gests that while this proportion is relatively small, individuals in this interventions provided by the AI system, leading to quicker age group may have different needs and expectations when it comes improvements. Some patients prefer face-to-face psychological to mental health services. Younger participants are typically more counseling or treatment and are less inclined to undergo interven- familiar with and confident in using new technologies, which results tion through the AI system. For these individuals, even if the AI in higher comfort and satisfaction when using such systems. In con- system offers personalized intervention plans, the effectiveness may trast, older participants may experience lower comfort and satisfac- be limited due to their lack of adaptation to or dislike for this tion due to a lack of experience with technology or fear of new approach. 15% of the respondents do not perceive a significant technologies. Regarding educational background, participants with a improvement in their mental health condition. Figure 6 illustrates high school education or below account for 27.78%, suggesting that the results. the IMHAS also reaches out to individuals with relatively lower levels of education. The highest proportion of participants holds an associate 4.4 Personalization assessment degree, accounting for 33.33%, indicating a potentially higher demand for mental health services among this educational group. Participants This work evaluates the level of personalization for the survey par- with a bachelor’s degree account for 22.22%, and they may have a ticipants. 85% of the respondents believe that the personalized inter- better understanding of and demand for mental health services. vention plans provided by the system are quite suitable for their actual Participants with a master’s degree or above constitute 16.67% of the situations and needs. 10% of the respondents consider the personal- total, and this group may have higher expectations for the quality of ized intervention plans provided by the system to be moderate. 5% of mental health services. the respondents find the personalized intervention plans provided by the system unsatisfactory. Figure 7 illustrates the results. 4.2 Satisfaction evaluation 4.5 Evaluation of interaction experience This work analyzes the satisfaction of the survey participants, with 75% of the respondents expressing satisfaction or high satisfaction This work evaluates the interaction experience of the survey par- with the IMHAS. 20% of the respondents indicate moderate satisfac- ticipants. 70% of the respondents report that the interaction with the tion with the system, while 5% express dissatisfaction. Figure 5 illus- system is smooth and natural. 20% of the respondents indicate that trates the results. The respondents mainly express dissatisfaction with there is room for improvement in the system’s recognition and under- the system’s level of personalization, ease of use, and the effectiveness standing of user emotional states. 10% of the respondents rate their of the interventions. They feel that the system’s personalized recom- interaction experience with the system as average. Figure 8 displays mendations are not accurate enough, the operational process is rela- the results. tively cumbersome, and the interventions do not lead to significant improvements in their condition. 4.6 Evaluation of data privacy protection This work assesses the survey participants’ perceptions of data privacy protection. Question 1: Do you believe the system effectively protects your personal privacy information? Average score: 4.2; Standard deviation: 0.8. Question 2: Are you concerned about your personal privacy information being leaked during the use of the system? Average score: 2.8; Standard deviation: 1.2. Question 3: Are you clear about how the system handles and protects your personal privacy information? Average score: 3.6; Standard deviation: 0.9. Figure 9 displays the results. For questions 1 and 3, the respondents’ evaluations are relatively consistent, with data points clustered around the mean. It indicates that the respondents’ evaluations of the system’s ability to effectively protect personal privacy information and the clarity of the system’s handling and protection of personal privacy information are relatively stable. For question 2, the respondents’ evaluations are relatively dis- FIGURE 5 Distribution of users across different levels of satisfaction. persed, with data points spread over a wider range. It suggests that there is some variability in the level of concern among respondents Frontiers in Psychology 09 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 FIGURE 6 FIGURE 8 Impact of the IMHAS on users’ mental health improvement. Users’ attitude towards interaction experience with the IMHAS. FIGURE 7 Users’ attitude towards personalized services provided by the IMHAS. FIGURE 9 Statistical evaluation of data privacy protection. about the potential leakage of personal privacy information during the use of the system, and the evaluations are not as consistent. is performed to verify the sphericity assumption, and the results show After the intervention, the mental health scale scores of the exper- that χ2 = 2.31, df = 1, p = 0.129 > 0.05, indicating that the sphericity imental group participants significantly increase, with noticeable hypothesis is valid and no Greenhouse–Geisser correction is required. reductions in anxiety and depression symptoms. A paired-samples Eighty-five percent of the experimental group participants believe t-test is adopted in this study to further compare the changes in that the personalized intervention plans provided by the IMHAS are PHQ-9 and GAD-7 scores before and after the intervention. The very suitable for them, 10% think they are moderately suitable, and results show that the average PHQ-9 score of participants in the exper- 5% think they are not suitable. Seventy percent of the participants in imental group decreased by 4.5 points (SD = 1.2), and the average the control group believe that traditional psychological intervention GAD-7 score decreased by 3.8 points (SD = 0.9). The paired-samples plans suit their needs, 20% think they are moderately suitable, and t-test confirms that these changes are statistically significant (PHQ-9: 10% think they are not suitable. The experimental group shows sig- t(179) = −15.2, p < 0.001, Cohen’s d = 1.12; GAD-7: t(179) = −12.3, nificantly better applicability of personalized intervention plans com- p < 0.001, Cohen’s d = 0.91). This indicates that the IMHAS system has pared to the control group (p < 0.05). a substantial effect on improving the mental health status of partici- pants. The Shapiro–Wilk test is conducted, and the results showed that 4.7 Psychological intervention strategies all continuous variables (PHQ-9 and GAD-7 scores) in both groups conformed to a normal distribution (p > 0.05), with no significant Analyzing the mental health data and behavioral data of social skewness or kurtosis anomalies. The Levene test verifies the homoge- assistance recipients and mining valuable information and patterns neity of variance for PHQ-9 and GAD-7 scores before and after the from them to provide a scientific basis for psychological intervention intervention in the two groups (p > 0.05), which meets the prerequi- is a work of potential significance. By analyzing the data obtained sites for the independent-samples t-test and analysis of variance. In from the questionnaire survey, the system can reveal the mental health the mixed-design analysis of variance, the Mauchly’s test of sphericity problems and risk factors that the recipients may have and propose Frontiers in Psychology 10 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 corresponding intervention suggestions and measures, as shown in recommendations, types of interventions, delivery methods, and treat- Table 1. ment outcomes. Among these, personalized recommendations and Participants generally express a desire for the system to be opti- treatment effectiveness are the two areas that participants are most mized and improved in areas such as personalized recommendations, concerned with. ease of use, and treatment effectiveness. These feedbacks provide clear A rigorous data governance framework is established in this study, direction and goals for future system enhancements. The above inter- including data encryption, anonymization and access control. All data vention suggestions and measures aim to provide scientific psycho- is encrypted with the Advanced Encryption Standard during collec- logical intervention support to social assistance recipients by analyzing tion, storage and processing to ensure security in data transmission psychological health data and behavioral data. The goal is to help them and storage. To enhance the transparency and interpretability of the improve their mental health status, enhance their quality of life, and algorithm, the development process of the AI system, as well as the achieve self-development and growth. Meanwhile, these measures also algorithms and model architectures adopted, are documented in require collaboration with professional mental health institutions and detail. A Data Security Emergency Response Plan is also formulated: professionals to jointly provide comprehensive mental health services in the event of a data breach, affected users would be notified within to the recipients. 24 h, with details on the scope of the breach, risk level and remedial measures. Meanwhile, the incident would be reported to the ethics review committee and data protection regulatory authorities, and cor- responding liabilities would be assumed. Multiple channels are used 5 Conclusion to recruit participants, including communities, medical institutions and online platforms, to ensure the diversity and representativeness of The survey conducted reveals that social assistance recipients gen- the sample. However, the principle of voluntary participation may lead erally express satisfaction or high satisfaction with the IMHAS. They to self-selection bias in the sample, i.e., individuals who pay more believe that the system effectively improves their mental health status attention to mental health issues or are more willing to accept new and provides personalized intervention plans. This indicates that the technological interventions may be more inclined to participate in the system has played an important role in social assistance, contributing study. Such bias may affect the generalizability and applicability of the to the enhancement of recipients’ mental health levels. Respondents research results. To mitigate this bias, random sampling methods will generally believe that the personalized intervention plans provided by be considered in subsequent studies, and efforts will be made to the system are in line with their actual situations and needs. This dem- expand the recruitment scope to attract more participants with diverse onstrates the significant importance of personalized intervention in backgrounds and needs. At the same time, sensitivity analysis will be mental health assistance, as it can better meet the needs of recipients conducted on the research results to evaluate the impact of different and improve intervention effectiveness. Participants show varying sample characteristics on the results, thereby improving the robust- degrees of satisfaction with the system’s personalized ness of the research conclusions. TABLE 1 Psychological intervention strategies for social assistance recipients. Different Specific measures dimensions Personalized psychological For those recipients who engage frequently in social networking activities but experience significant emotional fluctuations, training intervention plans courses on emotional regulation and stress management can be provided to help them better cope with stress and emotional fluctuations. A platform can be established for psychological counseling services, allowing recipients to access professional psychological counseling Provision of psychological support anytime, anywhere. Through analysis of social networking behavior data, recipients who may need psychological counseling counseling services support can be identified and timely assistance can be provided. Psychological counseling services can include various forms such as online chat, telephone counseling, and face-to-face counseling to meet the needs of different recipients. By analyzing social networking behavior data, the recipients’ social relationship networks can be identified to establish support networks, Establishment of support allowing recipients to feel more social and emotional support. Support networks can include different groups such as family members, networks friends, and colleagues, who can provide support and encouragement when recipients face difficulties or challenges. A tracking mechanism can be established to regularly assess and monitor the mental health status of recipients. Through continuous data Regular tracking and collection and analysis, changes and trends in mental health problems can be promptly identified, and intervention measures can be evaluation adjusted accordingly. Social networking platforms and other channels can be utilized to provide mental health education to recipients, enhancing their Provision of mental health awareness and understanding of mental health and improving their mental health literacy. Mental health education can include skills in education emotion management, stress relief methods, and interpersonal relationship handling, helping recipients establish positive and healthy psychological behavior habits. Community psychological health support groups can be established, allowing recipients to receive more support and assistance within Community support and the community. Through analysis of social networking data, recipients who may be suitable for participation in support groups can be participation identified and invited to join. Community support groups can provide emotional support, information support, and practical assistance, helping recipients better cope with challenges and difficulties in life. Frontiers in Psychology 11 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 Although this study reports the significant effectiveness of the Resources, Writing – original draft. YG: Validation, Visualization, IMHAS system in improving the mental health of social assistance Writing – original draft. beneficiaries, a cross-sectional research design with limited longitudi- nal follow-up is adopted. This design restricts the evaluation of the long-term sustainability of intervention effects and the accurate infer- ence of causal relationships. Future research should consider adopting Funding a longitudinal follow-up design with a longer duration and even con- ducting randomized controlled trials to more accurately assess the The author(s) declared that financial support was received for this sustainability of intervention effects and causal relationships. In addi- work and/or its publication. This work was supported by National tion, as participants took part on a voluntary basis, individuals who Social Science Foundation of China (Grant No. 24XGL036). pay more attention to mental health issues or are more willing to accept new technological interventions may be more inclined to par- ticipate in the study. Such self-selection bias may affect the generaliz- ability and applicability of the research results. Future research should Acknowledgments strive to expand the recruitment scope and adopt more random sam- pling methods to reduce selection bias. This study relies on self- Our heartfelt appreciation extends to the school teachers who reported measurement tools to assess mental health status. Although assisted in experiment administration. Gratitude is also extended to these tools are widely used and have certain validity in mental health all participants whose contributions enriched this study. research, they may still be affected by participants’ subjective bias. Future research should consider combining objective physiological indicators and more comprehensive assessment tools to improve the accuracy of measurements. Conflict of interest The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be Data availability statement construed as a potential conflict of interest. The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. Generative AI statement The author(s) declared that Generative AI was not used in the creation of this manuscript. Ethics statement Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial The studies involving humans were approved by the Academic intelligence and reasonable efforts have been made to ensure accuracy, Ethics Review Hubei University, China. The studies were conducted including review by the authors wherever possible. If you identify any in accordance with the local legislation and institutional requirements. issues, please contact us. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accor- dance with the national legislation and institutional requirements. Publisher’s note All claims expressed in this article are solely those of the authors Author contributions and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product YZ: Conceptualization, Data curation, Formal analysis, Writing – that may be evaluated in this article, or claim that may be made by its review & editing. WX: Methodology, Project administration, manufacturer, is not guaranteed or endorsed by the publisher. References Alam, A. (2022). Investigating sustainable education and positive psychology interven- initial referrals to psychosocial interventions in patients with serious mental illness. tions in schools towards achievement of sustainable happiness and wellbeing for 21st Epidemiol. Psychiatr. Sci. 33:e2. doi: 10.1017/S2045796024000015 century pedagogy and curriculum. ECS Trans. 107:19481. doi: 10.1149/10701.19481ecst Boyer, P. (2022). Why we blame victims, accuse witches, invent taboos, and invoke spirits: Alowais, S., Alghamdi, S., Alsuhebany, N., Alqahtani, T., Alshaya, A., Almohareb, S., et al. a model of strategic responses to misfortune. J. R. Anthropol. Inst. 28, 1345–1364. doi: (2023). Revolutionizing healthcare: the role of artificial intelligence in clinical practice. 10.1111/1467-9655.13826 BMC Med. Educ. 23:689. doi: 10.1186/s12909-023-04698-z Chen, Z. H., Lin, L., Wu, C. F., Li, C. F., Xu, R. H., and Sun, Y. (2021). Artificial intelligence Barbalat, G., Plasse, J., Chéreau-Boudet, I., Gouache, B., Legros-Lafarge, E., Massoubre, C., for assisting cancer diagnosis and treatment in the era of precision medicine. Cancer et al. (2024). Contribution of socio-demographic and clinical characteristics to predict Commun. 41, 1100–1115. doi: 10.1002/cac2.12215 Frontiers in Psychology 12 frontiersin.org Zhang et al. 10.3389/fpsyg.2026.1578545 Doraiswamy, P. M., Blease, C., and Bodner, K. (2020). Artificial intelligence and the future Parker, S., and Grote, G. (2019). Automation, algorithms, and beyond: why work design of psychiatry: insights from a global physician survey. Artif. Intell. Med. 102:101753. doi: matters more than ever in a digital world. Appl. Psychol. 71, 1171–1204. doi: 10.1111/ 10.1016/j.artmed.2019.101753 apps.12241 Fei, Z., Yang, E., Li, D. D.-U., Butler, S., Ijomah, W., Li, X., et al. (2020). Deep convolution Qiu, M., and Wu, Y. (2024). Understanding the experience of family caregivers of patients network based emotion analysis towards mental health care. Neurocomputing 388, with leukemia: a qualitative analysis of online blogs. Humanit. Soc. Sci. Commun. 11, 212–227. doi: 10.1016/j.neucom.2020.01.034 1–11. doi: 10.1057/s41599-024-02830-y Fu, Z., Burger, H., Arjadi, R., and Bockting, C. L. H. (2020). Effectiveness of digital psy- Rowe, J. P., and Lester, J. C. (2020). Artificial intelligence for personalized preventive chological interventions for mental health problems in low-income and middle-income adolescent healthcare. J. Adolesc. Health 67, S52–S58. doi: 10.1016/j. countries: a systematic review and meta-analysis. Lancet Psychiatry 7, 851–864. doi: jadohealth.2020.02.021 10.1016/s2215-0366(20)30256-x Saiyu, G., Shali, W., Yanfang, L., and Jingping, Z. (2022). Mental health of recipients after Goldberg, S. B., Flemotomos, N., Martinez, V. R., Tanana, M. J., Kuo, P. B., Pace, B. T., et al. cadaveric liver transplantation: a perspective from positive psychology. J. Clin. Nurs. 32, (2020). Machine learning and natural language processing in psychotherapy research: 4710–4718. doi: 10.1111/jocn.16530 Alliance as example use case. J. Couns. Psychol. 67, 438–448. doi: 10.1037/cou0000382 Sanderson, W. C., Arunagiri, V., Funk, A. P., Ginsburg, K. L., Krychiw, J. K., Gual-Montolio, P., Jaén, I., Martínez-Borba, V., Castilla, D., and Suso-Ribera, C. (2022). Limowski, A. R., et al. (2020). The nature and treatment of pandemic-related psy- Using artificial intelligence to enhance ongoing psychological interventions for emotional chological distress. J. Contemp. Psychother. 50, 251–263. doi: 10.1007/ problems in real- or close to real-time: a systematic review. Int. J. Environ. Res. Public s10879-020-09463-7 Health 19:7737. doi: 10.3390/ijerph19137737 Shao, R., He, P., Ling, B., Tan, L., Xu, L., Hou, Y., et al. (2020). Prevalence of depression Hou, T., Zhang, T., Cai, W., Song, X., Chen, A., Deng, G., et al. (2020). Social support and and anxiety and correlations between depression, anxiety, family functioning, social sup- mental health among health care workers during coronavirus disease 2019 outbreak: a port and coping styles among Chinese medical students. BMC Psychol. 8:38. doi: 10.1186/ moderated mediation model. PLoS One 15:e0233831. doi: 10.1371/journal.pone.0233831 s40359-020-00402-8 Jolly, P., Kong, D. T., and Kim, K. Y. (2020). Social support at work: an integrative review. Sloshower, J., Guss, J., Krause, R., Wallace, R., Williams, M., Reed, S., et al. (2019). J. Organ. Behav. 42, 229–251. doi: 10.1002/job.2485 Psilocybin-assisted therapy of major depressive disorder using acceptance and commit- Le Glaz, A., Haralambous, Y., Kim-Dufor, D. H., Lenca, P., Billot, R., Ryan, T. C., et al. ment therapy as a therapeutic frame. J. Context. Behav. Sci. 15, 12–19. doi: 10.1016/j. (2021). Machine learning and natural language processing in mental health: systematic jcbs.2019.11.002 review. J. Med. Internet Res. 23:e15708. doi: 10.2196/15708 33944788 Sun, W., Bocchini, P., and Davison, B. D. (2020). Applications of artificial intelligence for Lee, D., and Yoon, S. N. (2021). Application of artificial intelligence-based Technologies disaster management. Nat. Hazards 103, 2631–2689. doi: 10.1007/s11069-020-04124-3 in the Healthcare Industry: opportunities and challenges. Int. J. Environ. Res. Public Torous, J., Bucci, S., Bell, I. H., Kessing, L. V., Faurholt-Jepsen, M., Whelan, P., et al. Health 18:271. doi: 10.3390/ijerph18010271 (2021). The growing field of digital psychiatry: current evidence and the future of apps, Leung, Y. W., Wouterloot, E., Adikari, A., Hirst, G., de Silva, D., Wong, J., et al. (2021). social media, chatbots, and virtual reality. World Psych. 20, 318–335. doi: 10.1002/ Natural language processing-based virtual cofacilitator for online cancer support groups: wps.20883 protocol for an algorithm development and validation study. JMIR Res. Protoc. 10:e21453. Xu, L., Sanders, L., Li, K., and Chow, J. C. L. (2021). Chatbot for health care and oncology doi: 10.2196/21453 applications using artificial intelligence and machine learning: systematic review. JMIR Mishra, S. (2020). Social networks, social capital, social support and academic success in Cancer 7:e27850. doi: 10.2196/27850 higher education: a systematic review with a special focus on ‘underrepresented’ students. Zhang, M., Ding, H., Naumceska, M., and Zhang, Y. (2022). Virtual reality technol- Educ. Res. Rev. 29:100307. doi: 10.1016/j.edurev.2019.100307 ogy as an educational and intervention tool for children with autism Spectrum Mitchell, E. (2019). Negotiating vulnerability: the experience of long-term social security disorder: current perspectives and future directions. Behav. Sci. 12:138. doi: recipients. Sociol. Rev. 68:003802611987677. doi: 10.1177/0038026119876775 10.3390/bs12050138 Mylona, A., Avdi, E., and Paraskevopoulos, E. (2022). Alliance rupture and repair pro- Zhang, J., Wu, W., Zhao, X., and Zhang, W. (2020). Recommended psychological crisis cesses in psychoanalytic psychotherapy: multimodal in-session shifts from momentary intervention response to the 2019 novel coronavirus pneumonia outbreak in China: a failure to repair. Couns. Psychol. Q. 35, 814–841. doi: 10.1080/09515070.2021.2013162 model of West China hospital. Precision Clin. Med. 3, 3–8. doi: 10.1093/pcmedi/pbaa006 Frontiers in Psychology 13 frontiersin.org