AI IN PSYCHIATRIC DIAGNOSTICS – A REVIEW
Authors:
- Aleksandra SobieskaASAleksandra Sobieska
- Patrycja MularczykPMPatrycja Mularczyk
- Julia JaguszewskaJJJulia Jaguszewska
- Weronika Martyna GórskaWGWeronika Martyna Górska
- Mateusz OkońMOMateusz Okoń
- Alicja GawlikAGAlicja Gawlik
- Alicja KorpackaAKAlicja Korpacka
- Iga Maria Owczarek-DanowskaIOIga Maria Owczarek-Danowska
- Aleksandra WojteraAWAleksandra Wojtera
Open Access
- 2 July 2026
- journal article
- review article
- Published by RS Global Sp. z O.O. in International Journal of Innovative Technologies in Social Science
- Vol. 1 (3(51))
Abstract
Introduction: Artificial intelligence (AI) is increasingly being used in psychiatry, with side effects on solutions stemming from the subjectivity of diagnosis, limited care, and biological complexity, which is subject to threats. Mental disorders affect 293 million people worldwide and pose a burden on human health [9]. Aim: The aim of this review is to summarize the current state of knowledge on AI applications in psychiatric diagnostics, with specific focus on: (1) analysis of communication traffic of AI algorithms, (2) analysis of the results of AI-based primary control, (3) extension of methodological and ethical implications, and (4) extension of research. Methods: A review of the research literature was conducted in the field of Basic Language Processing (NLP) in digital phenotyping, AI-assisted neuroimaging, and the ethical and legal implications of implementing these technologies. Meta-analyses, specific reviews, and original empirical studies completed between 2015 and 2026 were analyzed. Results: A meta-analysis reported a cumulative AI diagnostic accuracy of 85% and a therapeutic efficacy of 84% in specific applications [8]. NLP enabled independent assessment, achieving an 86% (AUC 0.93) in studies on psychosis risk states [18]. Chatbots (Woebot, Wysa, Youper) demonstrate the consequences of problem occurrence and anxiety [9]. A review of 555 neuroimaging models revealed that 83.1% of the symptoms appear as a consequence rather than being triggered by a utility [32]. The most important ethical concerns were identified, including algorithm opacity ("black box"), liability, and data privacy [54, 56, 61]. Conclusions: AI in psychiatric diagnostics has demonstrated transformative potential, particularly in the areas of NLP and digital phenotyping, but current neuroimaging models require methodological improvements. The development of comprehensive ethical frameworks and extensions, simple algorithms, and model validation in large, population-based cohorts are essential. The ultimate success of AI in psychiatry will depend on striking a balance between technological innovation and respect for fundamental ethical values, while maintaining a paramount clinical role in diagnostic and therapeutic procedures.Keywords
This publication has 58 references indexed in Scilit:
- The Promise of Artificial Intelligence–Powered Speech Biomarkers in PsychiatryJAMA Network Open, 2026
- Multimodal Representation Learning for Parsing Biological Heterogeneity in Psychiatric NeuroimagingBiological Psychiatry, 2026
- Comparative diagnostic performance of artificial intelligence models in structural MRI for schizophrenia: A systematic review and meta-analysisAsian Journal of Psychiatry, 2025
- Neuroimaging and machine learning in eating disorders: a systematic reviewEating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity, 2025
- First-person disavowals of digital phenotyping and epistemic injustice in psychiatryMedicine, Health Care and Philosophy, 2023
- Artificial intelligence for biomarker discovery in Alzheimer's disease and dementiaAlzheimer's & Dementia, 2023
- Fairness of artificial intelligence in healthcare: review and recommendationsJapanese Journal of Radiology, 2023
- Evaluation of Risk of Bias in Neuroimaging-Based Artificial Intelligence Models for Psychiatric DiagnosisJAMA Network Open, 2023
- Artificial Intelligence and Chatbots in PsychiatryPsychiatric Quarterly, 2022
- Data-Driven Diagnostics and the Potential of Mobile Artificial Intelligence for Digital Therapeutic Phenotyping in Computational PsychiatryBiological Psychiatry: Cognitive Neuroscience and Neuroimaging, 2019