A support system that visualizes children's opinion expression and psychological state.

Abstract

In a system that provides psychological support based on dialogue with children, this technology offers support that enables information sharing (advocacy) that respects the child's wishes while appropriately understanding changes in the child's emotional state and the urgency of the situation. [Solution] The child support AI system according to the present invention comprises: dialogue history holding means for holding a history of conversations with a user; emotion analysis means for analyzing the user's emotional state based on input information included in the dialogue history; and state mark generation means for determining psychological urgency and generating a state mark based on two evaluation indicators: (i) the frequency (temporal density) of negative emotions appearing within a predetermined time period, and (ii) the emotional intensity in a single conversation. Furthermore, this system presents supporters with status markers corresponding to the determined psychological urgency, and also includes visualization means that, based on user permission to share, presents supporters with summarized information of the conversation content that has been permitted to be shared.

Description

Technical Field

[0001] The present invention relates to an AI support system in the field of child welfare and education support, which supports children to naturally express their emotions and opinions, and assists supporters to appropriately notice the psychological changes of children and realize continuous support.

Background Art

[0002] Conventionally, in child welfare facilities and educational settings, a method has been adopted where supporters grasp the psychological state from the words and actions of children and reflect it in individual support plans and the like. Also, through documents such as "Children's Rights Notes", support is provided for children to understand their rights and express them as necessary. However, these efforts mainly rely on paper media, and the mechanism for providing interactive and continuous emotional support is not fully developed. In particular, for children who cannot adequately verbalize their feelings or for children in an environment where it is difficult to build a relationship of trust, the hurdle of self-expression is high, and there are not a few cases where support does not reach them.

[0003] Article 12 of the Convention on the Rights of the Child stipulates the "right to freedom of expression", Article 3 stipulates "respect for the best interests", and Article 19 stipulates "protection from all forms of violence and abuse", but the mechanism for concretizing these in daily practice is insufficient.

[0004] Also, it has been pointed out that behind the serious delinquency and domestic troubles reported in recent years, there are emotional isolation and lack of opinion expression from early childhood, and the importance of early detection through daily conversations and building a reassuring relationship has been re-recognized.

[0005] Based on such a situation, the need for a mechanism to daily visualize children's emotions and opinions so that supporters do not miss changes and can appropriately respond in a continuous relationship is increasing.

Summary of the Invention

Problems to be Solved by the Invention

[0006] The problem that this invention aims to solve is to provide an environment in which children can safely express their emotions and opinions, and to enable supporters to appropriately understand and share that information, thereby realizing continuous and child-centered support. In particular, as a mechanism to concretely support the "right to express opinions" as stipulated in Article 12 of the Convention on the Rights of the Child, there is a need to create an environment in which even children who have difficulty verbalizing their emotions and thoughts can naturally express their inner selves through daily conversations, and to design a system in which they can safely communicate their feelings and desires.

[0007] Furthermore, the function of "advocacy," which involves understanding the feelings of children who are unable to express their opinions well and conveying them to supporters on their behalf, is also important. This invention aims for AI to play a part in this role, functioning as a supplementary means to appropriately convey children's voices to society. In addition, it has been pointed out in recent years that emotional isolation and difficulty in expressing one's will are behind serious delinquency and crimes committed by young people, and the importance of noticing such "silent signs" in everyday conversations is increasing.

[0008] This invention aims to help caregivers recognize a child's condition and provide appropriate support or opportunities for relationship repair by visualizing emotional changes and signs of anxiety at an early stage. Thus, providing a technical means that enables the visualization and communication of emotions is the problem that this invention aims to solve. [Means for solving the problem]

[0009] The present invention relates to a support system comprising: an emotion analysis means for maintaining a history of conversations with a user and analyzing emotions based on that history; a state mark generation means for generating a state mark indicating the user's psychological state based on the results of the emotion analysis; and a visualization means for visually presenting information such as the state mark to a supporter.

[0010] The AI ​​support system (100) according to the present invention provides a mechanism that makes it easy for a user (10) to naturally express their feelings, and enables a supporter (70) to understand the content of those feelings and provide continuous and appropriate support. In this invention, negative emotions are extracted from input information (including utterance information and nonverbal information) contained in the dialogue history, psychological urgency is determined based on the frequency of occurrence of such negative emotions (number of detections within a predetermined time; hereinafter also referred to as "temporal density"), and a status mark indicating the psychological urgency is output, thereby enabling the supporter to notice changes in the user's psychology early and provide appropriate support.

[0011] While the primary purpose of this invention is to provide support for continuously monitoring children's emotional states in child welfare facilities, school settings, and home environments, it is a support system that, as a psychological support technology based on emotion analysis and the utilization of dialogue history, possesses flexibility that makes it applicable to other welfare, education, and medical fields, as well as high industrial applicability.

[0012] This system (100) is designed to encourage natural emotional expression and the elicitation of opinions by allowing the user (10) to freely talk about the events and feelings of the day through everyday conversations with an AI that has a character. Flexible conversations tailored to the characteristics of the user (10) are possible through input means (20) that use multiple modalities such as voice input, text input, stamps, drawings, and even image input including facial recognition. In addition, the user (10) can visually confirm the emotional tendencies and state based on the content they have entered, which is also included in the configuration to encourage introspection and self-understanding. The generated state marks are visually output on a screen or the like through visualization means (60).

[0013] These input data are stored by the dialogue history retention means (30), and the emotion analysis means (40) performs natural language processing and emotion analysis to understand the user's (10) emotional tendencies and inner state. For example, even if the user says "I'm fine," it is possible to discern feelings of anxiety from their tone of voice, length of silence, and facial expressions.

[0014] The status mark generation means (50) generates status marks indicating the psychological state of the user (10) based on the results of the emotion analysis. In particular, in order to evaluate psychological urgency, it employs a so-called "dual logic" based on two evaluation indicators: when multiple negative emotions are detected within a certain period of time (i.e., when the temporal density of negative emotions is high) and when a high-intensity negative emotion is detected in a single conversation (i.e., when the emotional intensity is high). This makes it possible to accurately capture both the accumulation of continuous stress and sudden psychological crises and to present the supporter (70) with a status mark indicating an appropriate level of urgency.

[0015] The visualization means (60) is configured to visually present status markers, emotional change trends, and key points of the dialogue history to the supporter (70) in an easy-to-understand manner, enabling the supporter (70) to intuitively grasp the situation. Furthermore, the supporter (70) can communicate with the user (10) in a chat format through the visualization means (60), promoting tripartite communication between the user (10), the supporter (70), and the AI. It also includes an advocacy support function in which, based on the user's (10) permission, summarized information related to the dialogue content that has been permitted to be shared is displayed to the supporter (70) through the visualization means (60).

[0016] Furthermore, the support record and handover management system (80) has a configuration that maintains and manages records of past conversation history, support content, emotional changes, etc., in chronological order, so that support is not interrupted even when the supporter (70) changes. This allows the new supporter (70) to intuitively understand the user's (10) condition and provide continuous support.

[0017] Furthermore, the empathetic response means (90) generates and presents empathetic responses to the user (10) based on the results obtained by the emotion analysis means (40), and has a configuration that provides psychological reassurance through everyday conversation. In addition, this conversation history and emotion data are appropriately processed using anonymization methods such as differential privacy and k-anonymity, and stored in a database for continuous AI learning, contributing to the improvement of response quality.

[0018] The configuration according to the present invention makes it easier for the user (10) to express their emotions and opinions through natural dialogue, and enables the supporter (70) to make quick and accurate decisions regarding appropriate support, thereby establishing a continuous and child-centered psychological support system. [Effects of the Invention]

[0019] The present invention provides the following benefits. First, it provides an environment in which children can safely express their feelings through continuous interaction with an interactive AI. This ensures psychological safety and leads to emotional stability by allowing children to regularly talk to the AI ​​about emotions and anxieties that are difficult to put into words. In particular, for children who have difficulty expressing themselves in interpersonal relationships, the function of the AI ​​as a "supportive presence" is an important element that can provide emotional support.

[0020] Next, the present invention has a function that visualizes a child's emotions based on various input information such as voice, text, and images, and generates empathetic responses accordingly. This makes it easier for caregivers to intuitively understand a child's emotional state, and allows them to utilize changes in emotions as "support for empathy" while avoiding situations where children are disadvantaged by quantification or labeling.

[0021] Furthermore, in this invention, since the history of conversations with the AI ​​and emotional data are continuously accumulated, even if the support provider changes, it becomes possible to provide continuous and consistent support based on past information. This ensures that support is not interrupted, and a sense of security and trust is maintained for the child.

[0022] Furthermore, by devising prompt designs in the interaction between supporters and AI, the response quality of the AI itself can be improved, enabling optimal support for each individual child.

[0023] Through the collaboration between supporters and AI, the AI can more effectively elicit children's opinion expressions and emotional expressions, and generate delicate responses according to their mental states. As a result, the construction of a higher-quality support system is also a major advantage.

[0024] In this system, it is possible for the supporter to design and input the prompt by themselves to guide the conversation, and it is also possible to use in combination the configuration where the AI automatically generates a prompt according to the user's mental state and proposes it to the supporter. Thereby, the expertise of the supporter and the auxiliary function of the AI work in coordination, and more effective support is realized.

[0025] Thereby, the supporter can immediately make a flexible response according to the changes in the conversation content and mood, forming a three-way support environment for children, supporters, and AI. This not only guarantees children's opportunities to express their opinions but also contributes to the construction of a continuous trust relationship.

[0026] In addition, the results of children's emotional analysis and conversation content output by the AI may also serve as an opportunity for the supporter to reflect on their own ways of involvement. Especially in situations where there is a possibility that the supporter themselves may cause harm in the family environment, the visualization of children's inner changes by the AI can be an opportunity for early intervention from an ethical and welfare perspective.

[0027] Also, instead of making direct judgments or interventions, the AI functions as an auxiliary entity that presents signs of emotions, and provides a mechanism to support appropriate responses by promoting cooperation between third-party organizations or experts through the supporter.

[0028] Note that the present invention is premised on support through dialogue with AI, but has a fundamental design concept of avoiding excessive dependence on AI. AI is an auxiliary entity that presents signs of a child's emotions and opinions to the supporter, and the supporter himself / herself is responsible for the final judgment and response. In this way, even while the supporter and AI cooperate, the configuration is such that it does not impair the quality of human support that protects the dignity of the child.

[0029] Furthermore, through dialogue with AI, a child can express opinions in a natural form, and by recording and sharing this, it becomes possible to convey thoughts that could not be put into words to the supporter. This mechanism also guarantees the child's "right to express opinions" and promotes advocacy support, realizing a support system in which the child's voice is not left unheard.

[0030] Through these effects, the present invention realizes a mechanism in which a child can safely express his / her own voice and that voice is surely reflected in the support, and has high effectiveness as a new support tool in fields such as child welfare, education, and family. <00001 "02>

[0031] In this system, as an output format for status marks and emotional expressions, by using non-verbal and visual methods such as face icons, colors, animations, etc., intuitive support that does not depend on language becomes possible. Such a configuration can exhibit its usefulness even when this system is utilized in other language regions and different cultural environments in the future. On the other hand, since there are cultural differences in the interpretation of emotional expressions, it is desirable to cooperate with local experts, etc. and perform necessary adjustments and verifications when deploying internationally.

Brief Description of the Drawings

[0032] [Figure 1] It is a schematic configuration diagram of an AI support system according to an embodiment of the present invention.

Mode for Carrying Out the Invention

[0033] Figure 1 is a block diagram showing the configuration of an AI support system according to one embodiment of the present invention. In this figure, the user (child) (10), the AI ​​support system (100), and the supporter (70) exchange information with each other as a tripartite configuration, conceptually illustrating the flow of emotion input, analysis processing, state mark generation, and empathetic response / notification.

[0034] The present invention relates to an AI support system (100) that promotes continuous and appropriate support by analyzing emotions based on the dialogue history of a user (10), particularly a child or student, and visually presenting their psychological state to a supporter (70).

[0035] The AI ​​support system (100) comprises the following main components. The AI ​​support system (100) shown in Figure 1 is composed of an input means (20), a dialogue history retention means (30), an emotion analysis means (40), a state mark generation means (50), a visualization means (60), a support record / handover management means (80), a supporter (70), and an empathetic response means (90).

[0036] The dialogue history retention means (30) records and stores the content of daily conversations with the child. This includes voice input, text input, stamps and emojis, and preserves the user's (10) natural expressions.

[0037] The emotion analysis tool (40) analyzes the accumulated dialogue history and estimates the user's (10) emotional state from various data such as tone, speaking speed, intonation, keywords, and facial expressions (image recognition).

[0038] The status mark generation means (50) has a configuration that generates status marks indicating the psychological state of the user (10) based on the results of emotion analysis. Specifically, it evaluates psychological urgency using the frequency (temporal density) of negative emotions appearing within a predetermined time period as an indicator in the analyzed emotion data, and automatically outputs a status mark corresponding to the degree of urgency. For example, if negative emotions such as "lonely" or "sad" are detected multiple times within one hour, and the frequency (temporal density) of these emotions exceeds a predetermined threshold, it is assumed that there is an increase in psychological tension, and a high-urgency mark such as a red "HELP mark" is presented to the supporter (70).

[0039] In this invention, "statements containing dangerous words" refers to statements containing words related to the life, physical, or psychological safety of children. Specifically, this includes expressions such as "I want to die," "I want to disappear," "I've been hit," and "Help me." In response, when the status mark generation means (50) detects these phrases, it immediately sets the highest level of urgency without waiting for a determination based on the two evaluation indicators (frequency of occurrence (temporal density) and emotional intensity). This means that even a single statement indicating a serious crisis will be immediately presented to supporters without undergoing an assessment of its frequency of occurrence, thus preventing serious incidents from occurring due to oversight.

[0040] With these configurations, the status mark generation means (50) employs a dual logic structure based on two evaluation indicators to assess psychological urgency: the frequency (temporal density) of negative emotions appearing within a predetermined time period and the emotional intensity in a single conversation. This makes it possible to accurately capture both the accumulation of continuous stress and sudden crises, minimizing oversights, and to present appropriate urgency marks to supporters.

[0041] On the other hand, in cases where mild negative or positive emotions occur consecutively, the system is configured to output low-urgency marks such as a "mild mark" or a "stable mark" as signs of psychological change, based on the accumulation pattern of these emotions. This allows the supporter (70) to understand the daily emotional fluctuations and recovery trends, enabling them to use this information for continuous and appropriate support.

[0042] Thus, this configuration promotes awareness among supporters (70) and enables timely and appropriate support by presenting status marks in a stepwise and visual manner according to the presence or absence of psychological urgency.

[0043] The visualization means (60) presents generated status markers, emotional fluctuation trends, and dialogue history in a format that allows the caregiver (70) to understand, enabling them to intuitively grasp changes in the situation. Furthermore, the caregiver (70) can use the visualized information to communicate with the child via chat or voice, bridging the gap from AI-based interaction to real human relationships.

[0044] Furthermore, the dialogue content and sentiment data are anonymized and stored in a database for continuous AI learning, contributing to the discovery of similar cases and the improvement of dialogue patterns. Anonymization employs techniques such as differential privacy and k-anonymity, ensuring full consideration for privacy protection.

[0045] Furthermore, this invention is not intended for AI-based decision-making or automatic notification, but rather as a supplementary tool to help caregivers (70) notice changes in children. The information presented by the AI ​​is intended to help caregivers (70) make appropriate decisions, and the final response is to be made by the caregivers (70) or specialized organizations.

[0046] In this way, the present invention facilitates the expression of emotions that are difficult to verbalize, and realizes a supportive environment that maintains continuous relationships and psychological safety. Overall configuration of the embodiment: Application forms in 3 fields

[0047] This invention, as an "AI Children's Rights Notebook," visualizes children's own emotions and opinions, and connects them to appropriate support while building trusting relationships with caregivers and stakeholders.

[0048] Its scope of application is broad, and specific implementations are shown in the following three areas: the "child welfare" field in child welfare institutions and child independence support facilities, the "education" field linked to school education and the GIGA School Initiative, and the "family" field supporting relationships with parents and caregivers within the home.

[0049] All of these fields are based on the common technological idea of ​​"recording, visualizing, sharing, and supporting" children's emotions and conditions. By demonstrating individually optimized application examples in each field, we can showcase the flexibility and high industrial applicability of the present invention.

[0050] Furthermore, this invention is based on the premise that AI functions as a supportive presence for children in each field, and is designed to communicate changes in their condition and signs of their emotions to those involved. In this regard, the AI ​​does not directly make judgments or intervene, but rather functions as a supportive presence that encourages "awareness" and "introspection" so that those involved can take appropriate action. Examples in the field of child welfare

[0051] The AI-powered Children's Rights Notebook according to the present invention will be used in child welfare facilities (e.g., children's homes, child independence support facilities, mother-child support facilities, temporary shelters, etc.) to support the establishment of continuous and trustworthy relationships between children and supporters and to realize the protection of their rights.

[0052] Children residing in the facility can naturally express their daily emotions, concerns, and requests by regularly interacting with the AI ​​installed in this system. Children can choose a method that suits them, such as voice input, pictures, stamps, or chat format (text input), so even children who have difficulty with interpersonal communication can express themselves without difficulty.

[0053] This system uses AI to analyze emotional fluctuations and recurring patterns of difficulties based on data obtained through dialogue, making the child's psychological state visible to caregivers. Furthermore, if the child decides it's okay to share their requests or opinions (e.g., "I want more free time," "I'm anxious at night and can't sleep"), these are shared with caregivers in a summarized form. This mechanism creates an environment where children can confidently express their feelings, even about things they might find difficult to say directly.

[0054] Furthermore, based on emotion analysis, the AI ​​has a function that automatically generates a "status mark" indicating psychological urgency when it detects negative emotions that are repeated within a short period of time, or when it detects a single, highly intense negative statement, and displays it on the caregiver's management screen. In this way, the configuration, which determines psychological urgency using a dual logic, allows caregivers to recognize the child's deep-seated SOS early, leading to quick and appropriate support decisions.

[0055] Furthermore, notifications are made based on information such as emotional fluctuation trends, status indicators, and trend analysis, and care is taken to avoid infringing on the children's privacy.

[0056] Furthermore, for children who have entered child independence support facilities due to delinquency or other reasons, this system provides an opportunity to deepen self-understanding by encouraging introspection through dialogue with AI, allowing them to naturally verbalize their emotions and experiences. This dialogue also provides a sense of security to children who have difficulties in relationships with others, supporting the prevention of recidivism and the improvement of their social adaptation skills.

[0057] Furthermore, this system can be configured to automatically generate prompts (e.g., "You seem a little down today, would you like to talk to them?") based on the user's (child's) emotional analysis results and status markers, and present them to the caregiver to facilitate dialogue. This prompt generation method makes it easier for caregivers to respond flexibly according to the child's psychological state, contributing to the realization of continuous and effective support.

[0058] Furthermore, supporters can continuously adjust and develop the AI ​​by designing prompts while observing the AI's responses and emotional analysis output, which is expected to enable support tailored to the individual characteristics of each child. In this process, supporters themselves can also have an opportunity to reaffirm their relationship with the children and reflect on their responses.

[0059] This system aims to capture children's "voices" and promote support that respects them, and can even act as an advocate for children who have difficulty articulating their rights. Therefore, it can enhance the quality of child-centered care throughout the facility and become a new technological foundation supporting a culture of rights advocacy. Examples of implementation in the education sector:

[0060] This invention functions effectively in both learning support and emotional support in educational settings where children and students routinely use tablet devices and the like.

[0061] For example, students can interact with the AI ​​at their own convenience, such as after school or at home, to express their feelings and opinions about school life. The AI ​​analyzes emotions based on the input text and voice, and while continuing a conversation that is comfortable for the individual, it can also visualize emotional trends as needed and share them with teachers and staff.

[0062] Emotions are visualized using abstract "state marks" and "trend graphs," thus avoiding the disadvantages of labeling while allowing for an intuitive understanding of support needs and changes. Furthermore, if a student wishes to "share their opinion," the AI ​​summarizes the conversation and displays it on the teacher's management screen. This system ensures that the individual's voice is appropriately reflected in the support system, expands opportunities for expression, and also serves as an advocacy tool.

[0063] Furthermore, the information accumulated through interactions with the AI ​​will be used in collaboration with school counselors, school nurses, and other professionals, facilitating early detection and connection to multidisciplinary support. In this embodiment, a configuration is adopted in which psychological urgency is detected based on a dual logic that evaluates both the frequency of negative emotions appearing within a predetermined time (i.e., temporal density) and single, high-intensity negative expressions. In this process, analysis is performed based on the high frequency (temporal density) of specific negative emotions being repeated in a short period of time, making it possible to accurately detect the need for highly urgent support.

[0064] Furthermore, by incorporating a dashboard function that displays the emotional tendencies and psychological states of multiple students in a list format, teachers and staff can gain an overview of the support needs of the entire class. This configuration contributes to classroom management, the development of support plans by school teams, and the centralization of self-evaluation and record-keeping in the educational setting. Examples of implementation in the home economics field

[0065] The support system according to the present invention is designed to allow children to express the emotions they feel in their daily lives in a natural way at home. By converting and presenting changes and trends in these emotions into status marks that can be visually understood by guardians, it contributes to improving relationships within the family and alleviating emotional isolation.

[0066] For example, in dual-income households, if a child spends time alone after school or during meals, emotions such as "I was alone" or "I was lonely" will be repeatedly detected in the conversation with the AI. In particular, if the same type of negative emotion is detected multiple times (e.g., four or more times) within a predetermined time period (e.g., within one hour), and the temporal density exceeds a predetermined threshold, a status mark indicating "accumulation of loneliness" is generated as a highly urgent emotional state.

[0067] Furthermore, even if the evaluation criteria for frequency of occurrence (temporal density) are not met, for "single, strong negative statements" (e.g., "I want to die," "I've been hit," "I want to disappear"), the highest level of urgency is immediately set according to the intensity of the emotional word and the danger it implies, and this is immediately presented to the supporter. Thus, in this embodiment as well, a dual logic is employed that uses both frequency of occurrence (temporal density) and emotional intensity as criteria to evaluate psychological urgency, ensuring that serious psychological states are not overlooked.

[0068] This allows parents to non-intrusively notice changes in their child's inner world and provides an opportunity to consider restructuring their interactions with their child, such as "spending time together, even if only during dinner."

[0069] Furthermore, even in cases of scolding or physical contact within the home, if a child records that they were "scolded" or "hit," the emotional analysis results will be visualized as a negative status mark. This can prompt parents to reflect on whether they "might have been a little too harsh," potentially leading to actions that repair the relationship, such as saying "I'm sorry about earlier."

[0070] Furthermore, if signs of delinquent tendencies, such as rebellious attitudes, verbal abuse, apathy, or frequent silence, are found in the emotional patterns accumulated within the family, the AI ​​can present them to parents as "warning marks" or "risk marks" depending on their intensity and persistence. This function allows parents to notice these signs early, before they escalate into delinquency, and enables preventative support to be taken in cooperation with specialized institutions and schools.

[0071] This system is designed with the child's privacy and sense of security as the top priority in all its applications, and the AI-based information processing is solely intended to support relationships within the family and promote awareness. In other words, it is not designed to automatically notify or report the child's statements to a third-party organization in an indictment-like manner, but rather emphasizes monitoring and support for introspection.

[0072] In the future, with parental consent and as needed, it is envisioned that an API interface will be implemented to share some information, such as the child's status markers and emotional history, with third-party support organizations such as school counselors, local childcare support workers, and municipal child and family support centers. This will contribute to building a comprehensive support system that improves children's well-being not only within the family but also throughout the community. [Industrial applicability]

[0073] This invention can be used as a support system to visualize emotions and opinions and promote collaboration with stakeholders in various situations in children's daily lives, such as child welfare facilities, educational institutions, and home environments. In particular, the psychological support technology utilizing emotion analysis AI and dialogue history has high applicability as a next-generation support infrastructure in child welfare and education settings.

[0074] Furthermore, the dialogue history and emotion analysis data generated and stored by this invention may be used as a record of the child's growth and as reference information for future self-understanding, provided that consideration is given to the child's privacy and developmental stage, and with the consent of the guardian.

[0075] This type of information is expected to be used by parents to understand their child's long-term developmental process and to provide more individualized parenting support, thereby contributing to improving children's well-being from multiple perspectives.

[0076] The technology of this invention can be applied to other fields where support for emotions and communication is required. For example, in medical institutions, it is expected to contribute to the early detection and response to feelings of loneliness and anxiety in children undergoing long-term treatment, and in the field of counseling, it is expected to support deeper dialogue and the building of trust by visualizing emotional changes that are difficult to put into words.

[0077] Thus, this invention, as a new information technology that visualizes a person's internal information and promotes communication in support relationships, has broad industrial application potential, although its primary focus is on child welfare and education.

[0078] Furthermore, by utilizing the aforementioned visualization means, it is possible to grasp the trends in support situations across the entire facility and at the class level, and it can be used as a supplementary basis for formulating support plans, self-evaluation of educational and welfare institutions, and third-party evaluation. In this way, the present invention can play a role as a next-generation information support platform that contributes not only to improving the quality of individual support but also to improving the quality of organizational support. [Explanation of Symbols]

[0079] 10: User (child) 20: Input method 30: Means for maintaining dialogue history 40:Sentiment analysis means 50: Means for generating status marks 60: Visualization means 70:Supporter 80: Support Records and Handover Management Methods 90: Supportive and responsive means 100: AI support system (overall)

Claims

1. A dialogue history retention means (30) that retains the history of conversations with the user (10), A sentiment analysis means (40) that analyzes the user's emotional state based on the input information included in the dialogue history, A state mark generation means (50) that determines psychological urgency based on the results of the analysis of the aforementioned emotional state and generates a state mark, (i) the frequency (temporal density) of negative emotions occurring within a given time period, and (ii) Emotional intensity in a single dialogue A status mark generation means (50) that determines the psychological urgency based on the two evaluation indicators, A visualization means (60) presents the status mark corresponding to the determined psychological urgency to the supporter (70), and, based on the user's (10) permission to share, presents the supporter (70) with summarized information of the conversation content for which sharing has been permitted. A child support AI system characterized by comprising the following features.

2. In the child support AI system described in claim 1, If the status mark generation means (50) detects a phrase related to the life, physical, or psychological safety of a child, it will, regardless of the judgment result based on the two evaluation indicators, A child support AI system characterized by generating the status mark by setting the psychological urgency to the highest level.

3. In the child support AI system described in claim 1, The aforementioned input information includes audio information or text information and image information. The aforementioned emotion analysis means (40) The emotion inferred from the aforementioned audio or text information, Based on the emotions estimated from the aforementioned image information, A child support AI system characterized by analyzing the emotional state of the user.

4. In the child support AI system described in claim 1, The status mark generation means (50) is If a predetermined number of negative emotions are detected within the predetermined time, or If negative emotions exceeding a predetermined intensity threshold are detected, A child support AI system characterized by determining that the psychological urgency is high.

5. In the child support AI system according to any one of claims 1 to 4, Based on the emotional state analyzed by the emotion analysis means (40), A child support AI system further comprising empathetic response means (90) for generating empathetic responses to the user (10).

Images