DeepBrain AI Completes AI Teen Counselor Pilot, Cutting Prep Time by 80%

Domestic AI technology has presented a new solution to comfort the hearts of adolescents. Deepbrain AI has successfully completed field verification of an innovative 'Adolescent Psychological Counseling Kiosk' that combines AI human and emotion recognition technology. This is part of the 'Knowledge Service Industry Technology Development Project' promoted by the Ministry of Trade, Industry and Energy, with Deepbrain AI overseeing its planning and development. Amidst the unfortunate reality of rising adolescent suicide rates and deepening mental isolation, the need for easily accessible digital psychological counseling solutions is more urgent than ever. This kiosk was developed to meet such contemporary demands and consists of an AI kiosk for real-time psychological counseling for adolescents and an AI analysis system for administrators. The kiosk features an AI human counselor on screen who friendly greets adolescents and assists them in completing a questionnaire. In this process, the user's voice, facial expressions, and language are analyzed in real-time to precisely identify their emotional state. ETRI's (Electronics and Telecommunications Research Institute) advanced emotion recognition module supports this core technological capability. The collected data is automatically recorded in the administrator system, which completes the questionnaire through AI-based analysis and provides visualized emotional analysis results. This enables counselors to reduce their preparation time by nearly 80% and focus on clients based on more reliable data. The addition of emotional analysis results is also expected to significantly improve the reliability of the questionnaire data. Deepbrain AI has proven the excellence of its AI human technology through this verification. All major technical indicators, including PSNR (29.4) for video clarity, SSIM (0.90) for facial similarity, Landmark Distance (2.6) for facial expression recognition accuracy, and real-time response speed (1.3 seconds), recorded world-class levels. The emotion recognition accuracy using the ETRI module also showed results surpassing high-performance model standards, with non-verbal at 89.18%, para-verbal at 86.62%, and multimodal at 82.69%. In particular, a field verification targeting middle school students, conducted jointly with Yonsei University, recorded high satisfaction, receiving an average suitability score of 84.13 for the survey-based digital counseling scenario and an overall positive response rate of 87.3%, earning positive evaluations from adolescents. This suggests that AI can be an effective tool for adolescent psychological counseling. Deepbrain AI plans to gradually expand the distribution of this innovative service to various sites such as schools, youth welfare facilities, and local government counseling centers. Furthermore, it is preparing to leap into the global market through the development of multilingual versions and a transition to SaaS (Software as a Service). Deepbrain AI CEO Jang Se-young stated, “This project is an exemplary case where AI human and emotion recognition technology have made a practical contribution to solving social problems,” adding that the company plans to further expand the role of generative AI in various areas such as public counseling and welfare. Attention is focused on whether Deepbrain AI's efforts can provide warm comfort and practical help to adolescents in crisis.