[AI-Supported Systems in Mental Health Care: Digital Innovation or Loss of Human Care?]

Abstract

Abstract: The increasing implementation of AI-supported systems in mental healthcare raises complex ethical, clinical, and societal questions. This manuscript examines the opportunities, risks, and limitations of such technologies, taking into account relational process factors as well as clinical outcomes and real-world conditions of care delivery. AI-supported systems can help bridge waiting times, provide low-threshold psychoeducational interventions, promote motivation, symptom stabilization, and treatment engagement, and facilitate access for underserved or difficult-to-reach patient populations. They offer flexible, location- and time-independent support, strengthen self-efficacy, and may help reduce stigma-related barriers. At the same time, several risks must be considered: potential limitations in the therapeutic alliance between patients and therapists, parasocial substitution, increased social isolation, the emergence of a potential two-tier healthcare system, and ethical challenges associated with unforeseen deteriorations in patients' conditions. Evidence-based studies, including randomized controlled trials and meta-analyses, demonstrate moderate but consistent improvements in symptoms of depression, anxiety, and loneliness, as well as positive effects in resource-limited settings and waiting-list contexts. The manuscript emphasizes that the effectiveness of AI systems can only be fully realized through their integration with human therapeutic support, ethical reflection, and appropriate structural implementation conditions. The aim is a differentiated, evidence-based, and ethically informed integration in which AI systems complement rather than replace the therapeutic relationship, thereby ensuring equitable access to care, quality of therapeutic relationships, and sustainable health outcomes alike.