Recent breakthroughs in generative artificial intelligence and large multimodal models (LMMs) are driving significant transformations in education, research, and practice within psychiatric and mental health fields [1,2]. As the demand for early detection and continuous monitoring grows, the adoption of AI-especially in nursing settings facing chronic workforce shortages-is anticipated to accelerate [3-5]. Despite notable constraints and criticisms surrounding the use of AI in psychiatric and mental health nursing [6,7], generative and analytic AI systems that can process large-scale multimodal data-including text, speech, imaging, wearable-derived signals, and behavioral logsare increasingly recognized for their potential to enhance productivity and access. This includes applications in screening and intervention support, automation of documentation, and improvements in nursing education [2,5]. A diverse range of AI applications suitable for mental health nursing practice has been documented. The integration of AI-based technologies into mental health nursing is expected to facilitate more systematic detection, assessment, prediction, optimization, and recognition of patients' conditions and symptoms. Examples of these applications include machine-learning algorithms, natural language processing (NLP), digital phenotyping, and conversational agents designed to assist in the assessment, diagnosis, and treatment of mental health conditions [8]. Furthermore, the use of AI in mental health nursing education is expected to continue expanding. Long-standing challenges and limitations-including limited opportunities for clinical placement due to patient safety, legal, and ethical restrictions; increasing diversity in learner proficiency alongside insufficient personalization; and the difficulty of standardizing instruction in socio-emotional competencies such as empathy and therapeutic communication- may be mitigated through the strategic application of AI [9-12]. Examples of AI applications include: learning analytics that leverage log data, answer patterns, and response times to adapt task difficulty and provide individualized learning pathways; virtual standardized patients (V-SPs) and conversational agents that facilitate iterative practice of therapeutic communication and crisis intervention scenarios with immediate feedback; and multimodal simulations that integrate speech, facial expressions, and text to support near-real-time, rubric-based assessment of competencies such as empathy, active listening, clarification, and restatement [13-16]. Within mental health nursing research, the expansion of AI use in education and clinical practice-alongside AI-based analyses of large-scale datasets-has resulted in a rapidly growing body of findings [17,18]. To further solidify this trajectory, the evidence base for AI applications in mental health nursing should be strengthened through the systematic accumulation of robust, domain-specific data. However, considerations of safety and ethical use are imperative when applying AI to healthcare. The World Health Organization's latest guidance recommends expanding the use of large multimodal models (LMMs) only if appropriate safety and governance safeguards are in place [19]. Similarly, the International Council of Nurses (ICN), in its statement on 'Digital Health Transformation and Nursing Practice,' emphasizes the need to concurrently strengthen education, policy, and research to ensure that nurses employ technology safely and ethically [20]. In light of these developments and the demonstrated need for AI adoption in mental health nursing, the Journal of Korean Academy of Psychiatric and Mental Health Nursing (JKPMHN) has chosen the theme of its first Special Issue, "AI and Data-Driven Research in Mental Health Nursing," intending to feature a diverse range of studies on this topic. The Special Issue is scheduled for publication in November 2025, and the journal welcomes a broad array of contributions from researchers in the mental health nursing community.

The trajectory of AI adoption extends beyond research conduct; it is also accelerating within the peer-review processes of scholarly publishing [21]. Increasingly, reviewers are using AI as an assistive tool, which has led some to upload confidential manuscripts to publicly available AI systems, thereby violating core principles of peer review. Additionally, there have been attempts to exploit AI-assisted reviewing through prompt injection (i.e., embedding hidden instructions) and recurrent reports of safeguard-circumvention (jailbreak) vulnerabilities in large language models (LLMs). These practices pose a significant risk to the trustworthiness and rigor of peer review, warranting strong caution against indiscriminate reliance on AI during the review process [22,23].

Many journals and publishers are revising their policies to allow AI assistance only under strict conditions, such as disclosure, protection of confidentiality, and prior approval. Between 2023 and 2025, WAME, COPE, and the ICMJE issued similar guidelines stating that AI cannot be credited as an author, all AI usage must be transparently disclosed, and uploading confidential manuscripts to public AI systems is prohibited [24-26]. Similarly, JKPMHN has established policies indicating that AI cannot be listed as an author and that any AI usage must be clearly disclosed. Recently, JKPMHN has implemented additional policies that require disclosure of any AI use during the peer-review process and prohibit the unauthorized uploading of confidential manuscripts or review materials to public AI systems. Moving forward, JKPMHN will promote AI-related research in psychiatric and mental health nursing that prioritizes safety and ethics. At the same time, the journal will work to maintain the rigor and trustworthiness of peer review by establishing ethical principles governing AI use during the review process and will continue to pursue these objectives.