Elyoseph Zohar, Levkovitch Inbar, Haber Yuval, Levi-Belz Yossi
School of Counseling and Human Development, The Faculty Of Education, The University of Haifa, Haifa, Israel.
Department of Brain Sciences, Faculty of Medicine, Imperial College London, London, United Kingdom.
J Clin Psychiatry. 2025 Jun 30;86(3):24m15525. doi: 10.4088/JCP.24m15525.
Suicide risk assessment is a critical skill for mental health professionals (MHPs), yet traditional training in this area is often limited. This study examined the potential of generative artificial intelligence (GenAI)- based simulator to enhance self-efficacy in suicide risk assessment among MHPs. A quasiexperimental mixed methods study was conducted. Participants interacted with an AI-based simulator (AIBS) that embodied the role of a patient seeking suicide risk assessment. Each participant conducted a real-time risk assessment interview with the virtual patient and received comprehensive feedback on their assessment approach and performance. Quantitative data were collected through pre- and postintervention questionnaires measuring suicide risk assessment self efficacy and willingness to treat suicidal patients (using 11-point Likert scales). Qualitative data were gathered through open-ended questions analyzing participants' experiences, perceived benefits, and concerns regarding the AI simulator. Among the 43 participating MHPs, we found a significant increase in self efficacy scores from preintervention (mean = 6.0, SD = 2.4) to postintervention (mean = 6.4, SD = 2.1, < .05). Willingness to treat patients presenting suicide risk increased slightly from (mean = 4.76, SD =2.64) to (mean = 5.00, SD = 2.50) but did not reach significance. Participants reported positive experiences with the simulator, with high likelihood to recommend to colleagues (mean = 7.63, SD =2.27). Qualitative feedback indicated that participants found the simulator engaging and valuable for professional development. However, participants raised concerns about overreliance on AI and the need for human supervision during training. This preliminary study suggests that AIBSs show promise for improving MHPs' self-efficacy in suicide risk assessment. However, further research with larger samples and control groups is needed to confirm these findings and address ethical considerations surrounding AI use in suicide risk assessment training. AI powered simulation tools may have potential to increase access to training in mental health, potentially contributing to global suicide prevention efforts. However, their implementation should be carefully considered to ensure they complement rather than replace human expertise.
自杀风险评估是心理健康专业人员（MHPs）的一项关键技能，但该领域的传统培训往往有限。本研究考察了基于生成式人工智能（GenAI）的模拟器在提高心理健康专业人员自杀风险评估自我效能方面的潜力。进行了一项准实验性混合方法研究。参与者与一个体现寻求自杀风险评估患者角色的基于人工智能的模拟器（AIBS）进行互动。每位参与者与虚拟患者进行实时风险评估访谈，并收到关于其评估方法和表现的全面反馈。通过干预前后的问卷收集定量数据，这些问卷测量自杀风险评估自我效能和治疗自杀患者的意愿（使用11点李克特量表）。通过开放式问题收集定性数据，分析参与者对人工智能模拟器的体验、感知到的益处和担忧。在43名参与的心理健康专业人员中，我们发现自我效能得分从干预前（均值 = 6.0，标准差 = 2.4）到干预后（均值 = 6.4，标准差 = 2.1，p <.05）有显著提高。治疗有自杀风险患者的意愿从（均值 = 4.76，标准差 = 2.64）略有增加到（均值 = 5.00，标准差 = 2.50），但未达到显著水平。参与者报告了对模拟器的积极体验，很有可能向同事推荐（均值 = 7.63，标准差 = 2.27）。定性反馈表明，参与者发现模拟器引人入胜且对专业发展有价值。然而，参与者对训练期间过度依赖人工智能以及需要人类监督提出了担忧。这项初步研究表明，基于人工智能的模拟器在提高心理健康专业人员自杀风险评估自我效能方面显示出前景。然而，需要用更大的样本和对照组进行进一步研究，以证实这些发现并解决围绕在自杀风险评估培训中使用人工智能的伦理考量。人工智能驱动的模拟工具可能有潜力增加心理健康培训的可及性，有可能为全球自杀预防努力做出贡献。然而，应仔细考虑其实施，以确保它们补充而非取代人类专业知识。
