<![CDATA[Not revealing psychologically distressing thoughts, symptoms, and experiences, that is, psychological non-disclosure, contributes to the failure of detection and treatment in many cases of mental illness. Since diagnostic methods mostly depend on self-reporting, such states often remain unrecognized. This article is a literature review and it summarizes the research on how AI and computational methods can be applied to detect psychological non-disclosure and secret states. According to PRISMA 2020, literature was searched using three databases: Scopus, PubMed, and ScienceDirect; only eligible sources were selected and the Mixed Methods Appraisal Tool was used for evaluating. After excluding duplicates only 791 records remained with 20 of those (2020-2025) being included. Six methods are mentioned: language and text analysis; speech and acoustic biomarkers; facial micro-expressions and physiological signals; large language model conversational systems; predictive machine learning; and disclosure augmentation. The study shows that multimodal AI can identify hidden distress, self-injury, and suicidal states although the knowledge is still not fully developed, validation is not enough, and safeguards against misuse remain weak. A framework mapping the different modalities to possible targets was suggested as well the article highlights the need for more thorough validation, transparent reporting, and privacy-preserving use of AI.]]>
Copyrights © 2025
