LingxiDiagBench: A Multi-Agent Framework for Benchmarking LLMs in Chinese Psychiatric Consultation and Diagnosis
arXiv:2602.09379v3 Announce Type: replace-cross Abstract: Mental disorders are highly prevalent worldwide, but the shortage of psychiatrists and the inherent subjectivity of interview-based diagnosis create substantial barriers to timely and consistent mental-health assessment. Progress in AI-assisted psychiatric diagnosis is constrained by the absence of benchmarks that simultaneously provide realistic patient simulation, clinician-verified diagnostic labels, and support for dynamic multi-turn consultation. We present LingxiDiagBench, a large-scale multi-agent benchmark that evaluates LLMs on
The proliferation of powerful large language models and the increasing global demand for mental health services are converging, enabling AI to address critical gaps in psychiatric care and diagnosis.
This development indicates a significant step towards scalable, accessible, and potentially more consistent mental health assessment globally, leveraging AI to augment human medical expertise, particularly in regions with psychiatrist shortages.
The paradigm for mental health diagnostics is shifting towards AI-assisted consultation, moving beyond traditional interview-based methods and introducing a multi-agent framework for standardized and verifiable benchmarks.
- · AI developers (especially in healthcare)
- · Patients in underserved areas
- · Mental healthcare providers (efficiency gains)
- · China (as a leader in specific AI applications)
- · Traditional psychiatric assessment frameworks
- · Regions slow to adopt AI in healthcare
- · Manual data collection for mental health research
AI models become a standard tool for preliminary psychiatric assessment and consultation.
This leads to novel mental health care delivery models, potentially reducing the cost and increasing the availability of diagnostic services.
Ethical and regulatory frameworks become critical to manage the implications of AI-driven psychiatric diagnoses, particularly regarding privacy, bias, and accountability.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL