Artificial Intelligence in Mental Health: Benefits, Risks, and Ethical Considerations

Stigma has long shaped how mental health care is perceived and accessed. For example, Hirosawa et al. (2002) found that Japanese patients preferred changing the name of psychiatry departments due to stigma surrounding mental health care. As mental health care continues to develop, artificial intelligence (AI) is increasingly discussed as a tool that could support services in psychiatry and psychology.

AI has several potential benefits for mental health care. It can support clinicians by analysing patterns in data and helping with tasks such as early risk detection, progress monitoring, and personalized treatment recommendations. For instance, work on artificial emotional intelligence by Schuller and Schuller (2018) illustrates how AI can process emotional signals efficiently. In addition, Veldkamp (2023) describes psychometric AI as a way to use advanced data analytics to improve psychological measurement and, potentially, strengthen the quality and availability of mental health services.

AI is also being explored as a supportive tool in diagnosis. Khalifa and Albadawy (2024) examined MRI-based approaches using VSRAD for detecting depression and suggest that AI may help clinicians by adding extra diagnostic information, rather than replacing professional judgement.

Besides diagnosis, AI is influencing how care is delivered, especially through remote and digital formats. Remote therapy via augmented reality (AR) is still experimental, with limited clinical studies and relatively high costs. A systematic review by Bakır et al. (2023) examined AR use in mental health and describes it as an emerging area. Virtual reality (VR) exposure therapy and AR-based mindfulness tools are also early-stage innovations that sometimes include AI/ML features to strengthen therapeutic effects. Digital mental health platforms and text-based chatbots are often offered for free or through subscriptions and aim to provide scalable support. Audio- and video-based remote therapy, in contrast, is more established and widely used, often with flexible payment options, but it also brings legal and professional responsibility issues linked to remote care (Bakır et al., 2023). AI can also increase access in underserved regions: Shang et al. (2024) describe how AI-supported telehealth platforms and wearables can enable virtual consultations, track well-being, and support timely interventions in areas with limited specialist care.

However, these developments also create important challenges and ethical concerns. Because AI systems depend on large amounts of personal data, privacy and data security are major issues. The World Health Organization (2021) emphasizes that AI should be trained and tested on diverse datasets so it works fairly across different populations. Data quality is another key concern. Browning et al. (2024) point out that online recruitment can lead to poor-quality data, including fake survey responses generated by bots, which can distort results and increase bias. This is why strong validation checks are needed before AI tools are built or used in mental health contexts.

Engagement is also a challenge. Even when chatbots and conversational agents can improve access, many people stop using them quickly. Jabir et al. (2024) report high dropout rates in short-term studies and suggest ways to reduce attrition, such as combining chatbots with human support, using symptom tracking, and improving how the agent is presented (for example, visual features). This supports the idea that AI tools may be most effective when they complement human care rather than replace it.

Finally, AI can affect relationships and social life in ways that raise ethical questions. Virtual connections may provide companionship, but they can also involve risks such as manipulation, exploitation, or distancing from real-life relationships. For example, TOI Tech Desk (2024) reports increased interest in AI “boyfriends” among young women in China, reflecting growing acceptance of digital intimacy. In parallel, AI-driven simulations used in medical education may improve training and patient outcomes, but they still require clear ethical guidelines to prevent harm and ensure responsible use in research, teaching, and clinical practice.

Overall, AI may help expand and support mental health care, but only if ethical risks are managed carefully. Addressing privacy and data security, improving data quality, reducing bias, and protecting meaningful human care are essential for ensuring AI supports mental health services in a fair and responsible way (Poudel et al., 2025).

References:

Bakır, Ç. N., Abbas, S. O., Sever, E., Özcan Morey, A., Aslan Genç, H., & Mutluer, T. (2023). Use of augmented reality in mental health-related conditions: A systematic review. Digital Health, 9, 20552076231203649.

Browning, M. E., Satterfield, S. L., & Lloyd-Richardson, E. E. (2024). Mischievous responders: Data quality lessons learned in mental health research. Ethics & Behavior, 34(5), 303–313.

Guidance, W. H. O. (2021). Ethics and governance of artificial intelligence for health. World Health Organization, 1-165.

Hirosawa, M., Shimada, H., Fumimoto, H., Eto, K., & Arai, H. (2002). Response of Japanese patients to the change of department name for the psychiatric outpatient clinic in a university hospital. General Hospital Psychiatry, 24(4), 269–274.

Jabir, A. I., Lin, X., Martinengo, L., Sharp, G., Theng, Y. L., & Tudor Car, L. (2024). Attrition in conversational agent–delivered mental health interventions: Systematic review and meta-analysis. Journal of Medical Internet Research, 26, e48168.

Khalifa, M., & Albadawy, M. (2024). AI in diagnostic imaging: Revolutionising accuracy and efficiency. Computer Methods and Programs in Biomedicine Update, 5, 100146.

Poudel, U., Jakhar, S., Mohan, P., & Nepal, A. (2025). AI in mental health: A review of technological advancements and ethical issues in psychiatry. Issues in Mental Health Nursing, 1–9.

Schuller, D., & Schuller, B. W. (2018). The age of artificial emotional intelligence. Computer, 51(9), 38–46.

Shang, Z., Chauhan, V., Devi, K., & Patil, S. (2024). Artificial intelligence, the digital surgeon: Unravelling its emerging footprint in healthcare—The narrative review. Journal of Multidisciplinary Healthcare, 4011–4022.

Veldkamp, B. P. (2023). Trustworthy artificial intelligence in psychometrics. In Essays on Contemporary Psychometrics (pp. 69–87). Springer International Publishing.