Artificial Intelligence

Artificial intelligence (AI) has expanded rapidly across healthcare and mental health in recent years, fueled by the growth of electronic health record data, advances in computing power, and the widespread use of digital devices. In clinical medicine, AI applications now support diagnostics, risk prediction, imaging interpretation, and workflow optimization. In mental health, research and test programs have increased substantially in recent years. However, reviews of the evidence emphasize that many tools still require stronger validation, equity testing, and long-term evaluation.1,2

As these technologies evolve, their application has the potential to address gaps in maternal mental health screening, diagnosis, and access to care. The goal of this fact sheet is to provide a general overview of how AI is currently used in mental health, highlighting its potential, risks, and considerations.

AI in Healthcare: An Overview
  • Types of AI:
    • Generative AI: Creates content such as generating draft clinical notes, summarizing medical records, and drafting patient messages.
    • Machine Learning (ML): Used for diagnostics, predictive analytics, and analyzing medical images such as MRI/CT brain scans. 
    • Natural Language Processing (NLP): Processes and understands human language (voice and written form) used for clinical documentation, electronic medical record coding, and in-depth sentiment analysis of patient speech.
      • Includes Ambient Listening. Listens to sessions in real-time to automate documentation so providers can focus entirely on the patient when face-to-face. 
    • Conversational AI & Digital Phenotyping: Specialized chatbots delivering Cognitive Behavioral Therapy (CBT) and analyzing data from wearables to monitor mental health trends.
    • Robotic Process Automation (RPA): Handles repetitive, rule-based tasks like insurance claims processing, authorizations, and scheduling. 
    • Agentic AI: acts as an orchestrator between systems such as electronic health records and patient platforms. 
  • It is useful to classify AI tools used by healthcare providers by their functional role in care delivery, such as screening, monitoring, or treatment support, because regulatory oversight, ethical risks, and evaluation standards vary by clinical function rather than theoretical capability.3,4
Research Regarding AI in Mental Health 
  • In mental health, AI applications have become increasingly prominent in these areas:
    • Risk prediction and suicide modeling2,5
    • Symptom monitoring through digital phenotyping (smartphones and wearables)4,6
    • Natural language processing of clinical notes2,3
    • Chatbot-based psychological interventions1,7
  • There has been rapid growth in AI-based mental health research and pilot development utilizing AI, though many tools are still in early implementation stages.2–4       
  • According to a 2020 overview on AI use in mental health, AI is most commonly used for depression prediction and detection, followed by AI studies on schizophrenia, suicidal ideation/attempts, and general mental health. Studies used diverse data sources, including electronic health records, brain imaging, mood scales, smartphones, and social media.5
  • The performance of AI use in mental health varies widely depending on the data source and outcome. Reported prediction accuracies ranged from the low 60% range (e.g., smartphone or social media data) to the high 90% range (e.g., structured clinical and demographic data). Models also showed promise in predicting treatment response and medication adherence.5                    
  • Most AI research currently uses supervised machine-learning techniques, a method in machine learning where an algorithm learns from a labeled dataset to make predictions or decisions about new, unseen data, combined with natural language processing.5 
  • Ongoing concerns of use of AI include limited external validation, sample homogeneity, and unclear long-term effectiveness. AI chatbots showed inconsistent performance in treating mental health conditions.2 
  • Additional limitations identified in the use of AI to diagnose mental health conditions include:
    • The design of some AI-based diagnostic tools may be biased2
    • The inability of AI tools to establish causality2
    • The application of AI-assisted diagnosis includes trade-offs between different performance metrics (for instance, between model specificity and sensitivity)2
    • The limited generalizability of AI tools 2
Research Regarding AI in Maternal Mental Health 

AI in maternal mental health is currently concentrated on screening and risk prediction, using supervised machine-learning models applied to clinical, sociodemographic, and increasingly passive digital data to identify women at elevated risk of perinatal depression and anxiety.2,6,8 Research on AI and MMH is growing but still limited. A review on AI and MMH identified only 14 studies between 2013–2023, indicating that AI in perinatal mental health remains an emerging field.8

Research highlights include:

  • Currently, most studies focus on predicting postpartum depression using supervised machine-learning models to identify risk factors and predict postpartum depression using EHR, survey, or psychometric data.8
  • Important risk factors across studies include prior mental health history, obstetric complications, healthcare utilization, sociodemographic variables, relationship quality, and social support.8
  • Natural language processing has been used to detect symptoms, such as self-harm, in EHRs and conduct exploratory analysis on social media posts for paternal depression.8 
  • Voice assistants often provided incomplete or inappropriate clinical advice.8
  • Chatbots have been tested for CBT delivery and mood monitoring.8
  • A small randomized controlled trial comparing a perinatal-tailored chatbot with usual care found the chatbot was feasible and acceptable and associated with modest short-term improvements in symptom scores.7 
  • Major methodological and implementation gaps remain. Challenges with the use of AI in maternal mental health include limited external validation, data heterogeneity, and concerns about bias, privacy, and clinical readiness.8 
Conclusion 

AI is being utilized rapidly by the perinatal population and healthcare and mental health professionals.  It can and should be utilized responsibly by healthcare professionals to reduce administrative work, including automating routine tasks like appointment scheduling and reminders, summarizing talk therapy session notes, and automating coding and billing. AI holds the potential to improve diagnostic accuracy by reviewing medical records and other data, such as smartphone applications, to synthesize symptoms. 

References
  1. Firth, J., Torous, J., Nicholas, J., Carney, R., Rosenbaum, S., & Sarris, J. (2017). Can smartphone mental health interventions reduce symptoms of anxiety? A meta-analysis of randomized controlled trials. Journal of Affective Disorders, 218, 15–22. https://doi.org/10.1016/j.jad.2017.04.046 ↩︎
  2. Cruz-Gonzalez, P., He, A. W.-J., Lam, E. P., Ng, I. M. C., Li, M. W., Hou, R., Chan, J. N.-M., Sahni, Y., Vinas Guasch, N., Miller, T., Lau, B. W.-M., & Sánchez Vidaña, D. I. (2025). Artificial intelligence in mental health care: A systematic review of diagnosis, monitoring, and intervention applications. Psychological Medicine, 55, e18. https://doi.org/10.1017/S0033291724003295 ↩︎
  3. Shatte, A. B. R., Hutchinson, D. M., & Teague, S. J. (2019). Machine learning in mental health: A scoping review of methods and applications. Psychological Medicine, 49(9), 1426–1448. https://doi.org/10.1017/S0033291719000151 ↩︎
  4. Torous, J., Bucci, S., Bell, I. H., Kessing, L. V., Faurholt-Jepsen, M., Whelan, P., Carvalho, A. F., Keshavan, M., Linardon, J., & Firth, J. (2021). The growing field of digital psychiatry: Current evidence and the future of apps, social media, chatbots, and virtual reality. World Psychiatry, 20(3), 318–335. https://doi.org/10.1002/wps.20883 ↩︎
  5. Graham, S., Depp, C., Lee, E. E., Camille Nebeker, Xin Tu, Ho-Cheol Kim, & Dilip V. Jeste. (2019). Artificial Intelligence for Mental Health and Mental Illnesses: An Overview. Current Psychiatry Reports, 21(116). https://link.springer.com/article/10.1007/s11920-019-1094-0 ↩︎
  6. Bilal, A. M., Emma Fransson, Emma Bränn, Allison Eriksson, Mengyu Zhong, Karin Gidén, Ulf Elofsson, Cathrine Axfors, Alkistis Skalkidou, & Fotios C Papadopoulos. (2022). Predicting perinatal health outcomes using smartphone-based digital phenotyping and machine learning in a prospective Swedish cohort (Mom2B): Study protocol. BMJ Open, 12(4). https://bmjopen.bmj.com/content/12/4/e059033 ↩︎
  7. Suharwardy, S., Ramachandran, M., Leonard, S. A., Gunaseelan, A., Lyell, D. J., Darcy, A., Robinson, A., & Judy, A. (2023). Feasibility and impact of a mental health chatbot on postpartum mental health: A randomized controlled trial. AJOG Global Reports, 3(3), 100165. https://doi.org/10.1016/j.xagr.2023.100165 ↩︎
  8. Kwok, W. H., Zhang, Y., & Wang, G. (2024). Artificial intelligence in perinatal mental health research: A scoping review. Computers in Biology and Medicine, 177, 108685. https://doi.org/10.1016/j.compbiomed.2024.108685 ↩︎