Artificial Intelligence and OCD: Interactions in Symptoms, Diagnosis, Treatment, and Ethical Considerations

Abstract Purpose of Review Artificial intelligence (AI) and associated technologies, including machine learning, deep learning, and generative systems such as large language models, have rapidly expanded in capability, scale, and daily use in recent years. These systems are also being explored for applications in healthcare, including mental healthcare. For some mental health conditions, such as obsessive-compulsive disorder (OCD), aspects and features of AI technologies may have unintended and unexpected interactions and consequences. Herein, we review the current literature exploring how AI and OCD interact in daily life and treatment settings, and consider ethical concerns. Recent Findings We find literature on this topic to exist within several broad domains: (1) how AI is shaping OCD presentations and symptoms; (2) how AI is being used in diagnosis and prognosis; and, (3) how AI is being used in treatment. Legal and ethical implications of AI in mental healthcare span domains including: (1) Privacy, Confidentiality, and Autonomy; (2) Bias, Transparency, and Humanization; and, (3) Patient Safety and Clinical Oversight. Summary AI use in daily life and mental healthcare settings is accelerating and can impact lived experience with OCD. Clinicians are advised to assess for use of AI by patients, and invite a conversation about AI’s role in their patients’ lives. As AI systems are incorporated into treatment settings, we suggest maintaining a human-in-the-loop for oversight of these systems. Finally, ethical concerns should continue to be reviewed as systems advance and adapt, with particular attention to privacy, transparency, patient agency, legal and ethical responsibility, and provider accountability.