JD Health and Chongqing hospital build mental-health agent lab, release psychiatry knowledge-graph model

JD Health and the First Affiliated Hospital of Chongqing Medical University have signed a strategic AI healthcare agreement and will jointly build a psychological rehabilitation and medical agent laboratory. The signing took place at the third Conference on Interdisciplinary Integration and Innovation in Psychiatry, held in Chongqing under the theme of rebuilding mental-health services with AI; the China Medical Education Association hosted the event, with the hospital, a Chongqing Medical University adolescent health centre and JD Health as co-organisers.

Beyond the agreement, the two parties released a national mental-health knowledge-graph model and demonstrated a diagnostic model on stage, moving toward a product chain that runs from pre-consultation to diagnosis, treatment and recommendation. China National Radio's report frames the deal as another deep deployment of JD Health's AI-plus-specialty model.

The rollout plan is concrete. The datasets and model outputs will go online, and the partners will deploy them in primary and secondary schools in Chongqing and the wider alliance, with a plan to give 20,000 students free AI screening and assessment covering the full screening, diagnosis and treatment cycle.

A multicentre study on the validity of the adolescent screening model started at the same time. It uses a multicentre research design to test the tool's value in real clinical settings and to build evidence for earlier screening, identification and intervention. That study matters more than the lab plaque: it puts the product in front of clinical evidence rather than pilot enthusiasm.

The conference also saw the release of a white paper on how Chinese doctors' practice and learning behaviour are changing in the AI era. Based on platform data and a survey of 1,455 doctors, it covers learning, diagnosis and treatment, research, patient management and public education, and concludes that doctors are moving from single-point, offline, reactive work toward online, continuous, human-machine collaboration.

Doctor-facing tools were shown in an AI exhibition area. Zhiyi aggregates tens of millions of medical papers, guidelines and drug references to produce evidence summaries, knowledge graphs and topic reviews, while MedWork covers literature sorting, research design, data analysis and first-draft writing. A live roundtable demo captured experts' remarks in real time and produced a draft review that, after expert revision, could be submitted to a journal.

The applications rest on JD Health's existing mental-health network: more than 6,000 psychiatrists and 2,000 counsellors, specialist partnerships with several tertiary psychiatric hospitals, and a chain running from screening and assessment to diagnosis, medication and long-term follow-up. AI handles fast assessment and triage inside hospitals, while the company's internet hospital closes the loop from consultation to testing, prescribing and delivery outside them.

Mental-health care has long been constrained by scarce supply and broken follow-up, and the value of AI here is less about replacing doctors than absorbing high-frequency, low-yield work such as screening and follow-up. JD Health's hospital-by-hospital co-build approach is clear but labour-intensive, since every site needs its own adaptation. What to watch is whether the 20,000-student screening produces reusable evaluation data, and whether the multicentre results are strong enough to push these tools into routine care.