JMIR Mental Health

Research on artificial intelligence (AI) and mental health has focused largely on harms at deployment, including chatbot safety, sycophancy, and AI-associated delusions. Less attention has been paid to a prior question: whether the human-generated text and preference judgments that shape large language models are themselves clinically reliable, particularly when self-report may be distorted. This Viewpoint aims to develop the clinical psychiatric construct of collusion—the uncritical acceptance of an unreliable account—as an analytic lens for AI training and deployment, and to argue that the clinical reliability of training and preference data should be treated as an explicit trustworthy-AI criterion in mental-health–relevant systems. A conceptual synthesis of psychiatry, clinical psychology, and AI safety literature was undertaken. The analysis distinguishes three pipeline layers: pretraining corpora, preference data and posttraining methods, and deployment-time interaction. It maps the clinical construct of collusion against adjacent technical concepts, including sycophancy, reward overoptimization, grounding, refusal training, red-teaming, and live monitoring. The synthesis suggests that collusion-like dynamics are least applicable at the pretraining layer and most applicable at the preference-data and deployment layers, where unassessed user or labeler input can be reinforced without corroboration. Existing mitigations, including data curation, Constitutional AI, reward-model evaluation, grounded generation, refusal training, red-teaming, and postdeployment monitoring, address parts of this problem. However, these approaches are not yet organized around a clinically informed account of when self-report is unreliable. The central novelty is therefore not a generic claim about bias, but the proposal that clinical self-report reliability should be assessed as a distinct data-quality and governance dimension. Trustworthy-AI frameworks for mental-health–relevant applications should incorporate clinical expertise in self-report reliability into preference-data design, red-teaming, and postmarket surveillance. Adding the clinical reliability of training and preference data as an explicit criterion could complement existing technical safeguards while leaving empirical evaluation of clinician involvement as an open research agenda.

Common mental health disorders are highly prevalent among refugees. There is an urgent need to address the mental health burden in this population.

Artificial intelligence (AI)–based conversational tools are rapidly expanding within mental health care as a means of increasing access and scalability. At the same time, these systems introduce distinct safety risks arising from both user disclosures (eg, self-harm ideation) and inappropriate or inadequate AI responses.

Cognitive behaviorally based interventions have broad appeal and potential for impact when treating adult alcohol and other drug use. Digitally delivered cognitive behaviorally based interventions (dCBIs) may offer this impact with the benefit of increased accessibility. Although prior reviews have indicated the benefits of dCBIs on substance use outcomes, the extension to psychosocial functioning outcomes is unknown.

Positive aging, a concept found in positive psychology, serves as the theoretical foundation for this study. To age positively, one must manage hidden or unrecognized challenges, show flexibility in behavior and thought, adopt a positive outlook on problems involving regression, and make decisions that promote one’s well-being.

Digital mental health tools (DMHTs) offer scalable support, but engagement varies. Understanding the shapes of initiation and ongoing use is essential for effective design and implementation.

Large language models (LLMs) are poised to transform mental health care, offering advanced capabilities in diagnosis, prognosis, and decision support. Since their inception, numerous mental health-focused LLMs have emerged in the scientific literature, reflecting the growing interest in leveraging these models across various clinical applications. With a broad range of models available, diverse optimization strategies, and multiple use cases, reviewing the current landscape is critical to understanding where future impact lies.

Tardive dyskinesia (TD) is a common, often underrecognized movement disorder resulting from long-term antipsychotic use, yet its detection in routine mental health care remains inconsistent despite the availability of structured rating scales.

Digital mental health interventions (DMHIs) have been widely promoted to improve access to mental health care within the UK National Health Service (NHS), particularly following the COVID-19 pandemic. In 2015, a total of 48 technologies were reportedly used in NHS services in England, but over the past decade, substantial changes to regulatory requirements, evidence standards, and procurement processes have reshaped the digital mental health landscape. There is limited clarity regarding which DMHIs are currently being formally procured and funded by NHS mental health services across the United Kingdom.

Major depressive disorder (MDD) affects approximately 1 in 6 adults during their lifetime, yet antidepressant selection relies predominantly on trial-and-error, with response rates of only 42% to 53%. While machine learning (ML) models have shown promise in predicting treatment outcomes, most focus on single treatments rather than comparative selection across therapeutic alternatives, limiting their clinical utility for the medication choice decisions that clinicians face in practice.

Problematic use of digital technology has increased across the world. Despite growing research, evidence on treatment effectiveness across digital behaviors remains fragmented.

Well-being is a cornerstone of public health and social progress; yet, its determinants are multifaceted and dynamic. As behavioral data become increasingly available and artificial intelligence (AI) systems gain prominence, scalable assessments of well-being are becoming more feasible. However, to be useful in practice, such systems must remain understandable to the people they aim to support. Explainable AI is therefore essential to foster trust and enable reflection.