{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,11]],"date-time":"2026-09-11T10:41:55Z","timestamp":1789123315627,"version":"build-2803163510"},"reference-count":56,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,11]],"date-time":"2026-09-11T00:00:00Z","timestamp":1789084800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Psychiatry"],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>General-purpose large language models are increasingly used by patients and caregivers to obtain mental health information and guidance about when professional care is required. In late-life depression, broadly accurate information may nevertheless be unsafe when cognitive change, multimorbidity, frailty, polypharmacy, self-neglect, caregiver dependence, or suicide risk is not adequately recognized. This study compared the clinical accuracy, safety, geriatric-specific appropriateness, triage performance, and response consistency of three large language models when answering patient- and caregiver-centered questions about late-life depression.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We conducted a blinded, paired benchmarking study using 90 questions covering six geriatric psychiatry domains and equally distributed across low-, moderate-, and high-risk strata. Each question was submitted independently to GPT-5.5 Instant via ChatGPT, Gemini 3.5 Flash via Gemini, and Seed2.0 Pro via Doubao, generating 270 primary-round responses. A stratified subset of 30 questions was resubmitted in separate conversations to assess test-retest consistency, yielding 360 responses overall. Two psychiatrists independently evaluated anonymized outputs against prespecified item-specific reference standards, with clinically important disagreements adjudicated by a third senior psychiatrist. The primary outcome was the proportion of clinically acceptable responses, defined using accuracy, clinical safety, geriatric appropriateness, warning-sign recognition, and triage criteria.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      Clinically acceptable responses were generated for 78.9% of questions by ChatGPT, 72.2% by Gemini, and 60.0% by Doubao (overall P&amp;lt;0.001). The difference between ChatGPT and Gemini was not statistically significant, whereas both outperformed Doubao. This model ranking remained unchanged under alternative core-safety and more stringent optimal-response definitions. Complete geriatric-specific appropriateness was achieved in 64.4%, 55.6%, and 43.3% of responses, respectively. Major safety errors occurred in 5.6% of ChatGPT responses, 10.0% of Gemini responses, and 16.7% of Doubao responses (raw P = 0.015; FDR-adjusted q=0.023). Performance declined substantially with increasing clinical risk. In exploratory\n                      <jats:italic>post hoc<\/jats:italic>\n                      caregiver-centered analyses, clinical acceptability was 76.7% for ChatGPT, 70.0% for Gemini, and 53.3% for Doubao, while explicit caregiver-directed action was present in 80.0%, 70.0%, and 53.3% of responses, respectively. Among high-risk questions, clinically acceptable response rates were 63.3%, 50.0%, and 36.7%, while under-triage occurred in 16.7%, 26.7%, and 36.7% of responses, respectively. Test-retest clinical consistency was highest for ChatGPT (90.0%), followed by Gemini (83.3%) and Doubao (73.3%).\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>The three models answered many late-life depression questions accurately and safely, but none demonstrated consistently reliable performance across complex or high-risk scenarios. Clinically important weaknesses involved geriatric-specific interpretation, recognition of indirect risk signals, crisis-response completeness, under-triage, and response stability. General-purpose large language models may support selected low-risk educational tasks but should not independently guide emergency triage, medication changes, suicide-risk management, or other safety-critical decisions in older adults.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fpsyt.2026.1956736","type":"journal-article","created":{"date-parts":[[2026,9,11]],"date-time":"2026-09-11T09:59:24Z","timestamp":1789120764000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Large language models for late-life depression: a blinded benchmark of clinical safety, geriatric appropriateness, and triage"],"prefix":"10.3389","volume":"17","author":[{"given":"Wei","family":"Xiao","sequence":"first","affiliation":[{"name":"Department of Plastic and Aesthetic Surgery, The Second Affiliated Hospital of Guilin Medical University","place":["Guilin, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Huanyu","family":"Zhang","sequence":"additional","affiliation":[{"name":"Tourism College, Beijing Union University","place":["Beijing, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiaoyi","family":"Chen","sequence":"additional","affiliation":[{"name":"School of Clinical Medicine, Xiangnan University","place":["Chenzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun","family":"Cai","sequence":"additional","affiliation":[{"name":"Department of Plastic and Aesthetic Surgery, The Second Affiliated Hospital of Guilin Medical University","place":["Guilin, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xuchen","family":"Luo","sequence":"additional","affiliation":[{"name":"Department of Plastic and Aesthetic Surgery, The Second Affiliated Hospital of Guilin Medical University","place":["Guilin, China"]},{"name":"School of Economics, Zhejiang University of Science & Technology","place":["Hangzhou, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiehua","family":"Deng","sequence":"additional","affiliation":[{"name":"Department of Plastic and Aesthetic Surgery, The Second Affiliated Hospital of Guilin Medical University","place":["Guilin, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,11]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1016\/j.pop.2022.10.010","article-title":"Geriatric depression","volume":"50","author":"Gundersen","year":"2023","journal-title":"Primary Care"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.1080\/13607863.2019.1699900","article-title":"Clinical characteristics of late-life depression predicting mortality","volume":"25","author":"van den Berg","year":"2021","journal-title":"Aging Ment Health"},{"key":"B3","doi-asserted-by":"publisher","DOI":"10.4102\/safp.v64i1.5534","article-title":"Late-life depression and the family physician","volume":"64","author":"Moodley","year":"2022","journal-title":"South Afr Family Pract Off J South Afr Acad Family Practice\/primary Care"},{"key":"B4","first-page":"51","article-title":"Late-life depression, clinical picture, etiology, therapeutic options","volume":"27","author":"Belteczki","year":"2025","journal-title":"Neuropsychopharmacologia Hungarica A Magyar Pszichofarmakologiai Egyesulet Lapja = Off J Hungarian Assoc Psychopharmacol"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.1080\/07317115.2021.1877912","article-title":"Recent research on late life depression","volume":"44","author":"Moye","year":"2021","journal-title":"Clin Gerontologist"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.1176\/appi.ajp.20230392","article-title":"Cognitive outcomes after antidepressant pharmacotherapy for late-life depression: A systematic review and meta-analysis","volume":"181","author":"Ainsworth","year":"2024","journal-title":"Am J Psychiatry"},{"key":"B7","article-title":"Delirium in older persons: Prevention, evaluation, and management","volume":"108","author":"Jaqua","year":"2023","journal-title":"Am Family Physician"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.1177\/07334648221146770","article-title":"The risk for loneliness and major depression among solo agers","volume":"42","author":"Adams","year":"2023","journal-title":"J Appl Gerontology Off J South Gerontological Soc"},{"key":"B9","doi-asserted-by":"publisher","first-page":"7","DOI":"10.1176\/appi.ajp.20230919","article-title":"Optimizing the treatment of late-life depression","volume":"181","author":"Walaszek","year":"2024","journal-title":"Am J Psychiatry"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1080\/14656566.2021.1921736","article-title":"An update on antidepressant pharmacotherapy in late-life depression","volume":"22","author":"Brender","year":"2021","journal-title":"Expert Opin Pharmacother"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1016\/j.jad.2026.121214","article-title":"Injurious falls before, during, and after depression diagnosis: Results from two longitudinal population-based studies","volume":"400","author":"Zhang","year":"2026","journal-title":"J Affect Disord"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.1093\/ageing\/afaf182","article-title":"Targeting oral frailty indicators of late-life cognitive disorders and depression: A systematic review","volume":"54","author":"Dibello","year":"2025","journal-title":"Age Ageing"},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.1177\/09567976241279203","article-title":"The well-being costs of informal caregiving","volume":"35","author":"Kr\u00e4mer","year":"2024","journal-title":"psychol Sci"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1177\/15333175211064756","article-title":"Perceived stress in older dementia caregivers: Mediation by loneliness and depression","volume":"37","author":"Peavy","year":"2022","journal-title":"Am J Alzheimer's Dis Other Dementias"},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1097\/YCO.0000000000001009","article-title":"Suicide in later life: The role of frailty and depression","volume":"38","author":"Draper","year":"2025","journal-title":"Curr Opin Psychiatry"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1017\/neu.2024.50","article-title":"Use of generative artificial intelligence (AI) in psychiatry and mental health care: A systematic review","volume":"37","author":"Kolding","year":"2024","journal-title":"Acta Neuropsychiatr"},{"key":"B17","doi-asserted-by":"publisher","first-page":"45","DOI":"10.1186\/s12929-025-01131-z","article-title":"Impact of large language model (ChatGPT) in healthcare: An umbrella review and evidence synthesis","volume":"32","author":"Iqbal","year":"2025","journal-title":"J BioMed Sci"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1177\/08901171251329650","article-title":"Should my chatbot health coach be certified and regulated","volume":"39","author":"Terry","year":"2025","journal-title":"Am J Health Promotion AJHP"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.1016\/j.arthro.2024.08.029","article-title":"Editorial commentary: Large language models like ChatGPT show promise, but clinical use of artificial intelligence requires physician partnership","volume":"41","author":"Ramkumar","year":"2025","journal-title":"Arthroscopy J Arthroscopic Related Surg Off Publ Arthroscopy Assoc North America Int Arthroscopy Assoc"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.1038\/s41582-022-00698-7","article-title":"The inter-relationship between delirium and dementia: The importance of delirium prevention","volume":"18","author":"Fong","year":"2022","journal-title":"Nat Rev Neurol"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2026.4003","article-title":"Large language model performance and clinical reasoning tasks","volume":"9","author":"Rao","year":"2026","journal-title":"JAMA Netw Open"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-026-04431-5","article-title":"General-purpose large language models outperform specialized clinical AI tools on medical benchmarks","volume":"32","author":"Vishwanath","year":"2026","journal-title":"Nat Med"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.1007\/s00115-023-01444-0","article-title":"Late-life depression and frailty-epidemiological, clinical and neurobiological associations","volume":"94","author":"Depping","year":"2023","journal-title":"Der Nervenarzt"},{"key":"B24","doi-asserted-by":"publisher","DOI":"10.1080\/13607863.2020.1857689","article-title":"Depression as a determinant of frailty in late life","volume":"25","author":"Borges","year":"2021","journal-title":"Aging Ment Health"},{"key":"B25","doi-asserted-by":"publisher","DOI":"10.1097\/ANA.0000000000000971","article-title":"Postoperative delirium and the older adult: Untangling the confusion","volume":"36","author":"Thedim","year":"2024","journal-title":"J Neurosurgical Anesthesiology"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.1016\/j.ajp.2022.103213","article-title":"Suicidality in the geriatric population","volume":"75","author":"Chattun","year":"2022","journal-title":"Asian J Psychiatry"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.1080\/07317115.2022.2039827","article-title":"Suicide risk among older adults: The moderating role of hope","volume":"47","author":"McLaren","year":"2024","journal-title":"Clin Gerontologist"},{"key":"B28","doi-asserted-by":"publisher","first-page":"25","DOI":"10.1016\/bs.pbr.2023.02.004","article-title":"Treatment resistant depression in elderly","volume":"281","author":"Wang","year":"2023","journal-title":"Prog Brain Res"},{"key":"B29","doi-asserted-by":"publisher","first-page":"31652","DOI":"10.1038\/s41598-025-17242-4","article-title":"Performance of mental health chatbot agents in detecting and managing suicidal ideation","volume":"15","author":"Pichowicz","year":"2025","journal-title":"Sci Rep"},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.2196\/69709","article-title":"A comparison of responses from human therapists and large language model-based chatbots to assess therapeutic communication: Mixed methods study","volume":"12","author":"Scholich","year":"2025","journal-title":"JMIR Ment Health"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1016\/S2215-0366(25)00396-7","article-title":"Artificial intelligence-associated delusions and large language models: Risks, mechanisms of delusion co-creation, and safeguarding strategies","volume":"13","author":"Morrin","year":"2026","journal-title":"Lancet Psychiatry"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.1080\/17482631.2026.2621802","article-title":"The benefits and future potential of generative artificial intelligence (gai) on mental health: A delphi study","volume":"21","author":"Oo","year":"2026","journal-title":"Int J Qual Stud Health Well-being"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiolchem.2025.108611","article-title":"Generative artificial intelligence and large language models in smart healthcare applications: Current status and future perspectives","volume":"120","author":"Haque","year":"2026","journal-title":"Comput Biol Chem"},{"key":"B34","doi-asserted-by":"publisher","DOI":"10.1097\/JS9.0000000000003668","article-title":"Current concerns and future directions of large language model ChatGPT in medicine: A machine-learning-driven global-scale bibliometric analysis","volume":"112","author":"Guo","year":"2026","journal-title":"Int J Surg (London England)"},{"key":"B35","doi-asserted-by":"publisher","first-page":"641","DOI":"10.1007\/s00345-025-06019-z","article-title":"Large language model chatbots for patient education in kidney stones: A scoping review","volume":"43","author":"Goudrar","year":"2025","journal-title":"World J Urol"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1016\/j.psc.2023.02.008","article-title":"Treatment-resistant late-life depression: A review of clinical features, neuropsychology, neurobiology, and treatment","volume":"46","author":"Subramanian","year":"2023","journal-title":"Psychiatr Clinics North America"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.2196\/70436","article-title":"Effectiveness of topic-based chatbots on mental health self-care and mental well-being: Randomized controlled trial","volume":"27","author":"Tong","year":"2025","journal-title":"J Med Internet Res"},{"key":"B38","doi-asserted-by":"publisher","DOI":"10.1097\/ICO.0000000000003747","article-title":"Use of online large language model chatbots in cornea clinics","volume":"44","author":"Nichani","year":"2024","journal-title":"Cornea"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.1016\/j.arr.2022.101803","article-title":"Prediction models for depression risk among older adults: Systematic review and critical appraisal","volume":"83","author":"Tan","year":"2023","journal-title":"Ageing Res Rev"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.1093\/arclin\/acac083","article-title":"Reliable cognitive decline in late-life major depression","volume":"38","author":"Manning","year":"2023","journal-title":"Arch Clin Neuropsychol Off J Natl Acad Neuropsychologists"},{"key":"B41","doi-asserted-by":"publisher","first-page":"160","DOI":"10.1038\/s41398-023-02464-9","article-title":"Biological factors influencing depression in later life: Role of aging processes and treatment implications","volume":"13","author":"Szymkowicz","year":"2023","journal-title":"Transl Psychiatry"},{"key":"B42","doi-asserted-by":"publisher","first-page":"97","DOI":"10.1016\/j.gerinurse.2023.10.026","article-title":"Internet-delivered psychological interventions for older adults with depression: A scoping review","volume":"55","author":"Peng","year":"2024","journal-title":"Geriatric Nurs (New York NY)"},{"key":"B43","doi-asserted-by":"publisher","DOI":"10.1017\/S1041610223000066","article-title":"Neuroimaging features of depression-frailty phenotype in older adults: A pilot study","volume":"35","author":"Shuster","year":"2023","journal-title":"Int Psychogeriatrics"},{"key":"B44","doi-asserted-by":"publisher","first-page":"777","DOI":"10.1016\/j.jad.2025.03.150","article-title":"Association between loneliness and suicidality among general populations and persons with depressive and bipolar disorders: A systematic review","volume":"380","author":"Lu","year":"2025","journal-title":"J Affect Disord"},{"key":"B45","doi-asserted-by":"publisher","first-page":"22","DOI":"10.1093\/ijnp\/pyaa068","article-title":"Esketamine nasal spray for rapid reduction of depressive symptoms in patients with major depressive disorder who have active suicide ideation with intent: Results of a phase 3, double-blind, randomized study (ASPIRE II)","volume":"24","author":"Ionescu","year":"2021","journal-title":"Int J Neuropsychopharmacol"},{"key":"B46","doi-asserted-by":"publisher","DOI":"10.1016\/j.arr.2025.102710","article-title":"Systematic review of clinical effectiveness of interventions for treatment resistant late-life depression","volume":"107","author":"Pozuelo Moyano","year":"2025","journal-title":"Ageing Res Rev"},{"key":"B47","doi-asserted-by":"publisher","DOI":"10.1016\/j.neurom.2021.10.019","article-title":"Differential placebo responses for pharmacotherapy and neurostimulation in late-life depression","volume":"26","author":"Wathra","year":"2023","journal-title":"Neuromodulation J Int Neuromodulation Soc"},{"key":"B48","doi-asserted-by":"publisher","DOI":"10.1097\/YCO.0000000000001022","article-title":"Understanding late-life depression: Focus on inflammation","volume":"38","author":"Teixeira","year":"2025","journal-title":"Curr Opin Psychiatry"},{"key":"B49","doi-asserted-by":"publisher","DOI":"10.3390\/ijms23010308","article-title":"Pathomechanisms of vascular depression in older adults","volume":"23","author":"Jellinger","year":"2021","journal-title":"Int J Mol Sci"},{"key":"B50","doi-asserted-by":"publisher","DOI":"10.1136\/bmj-2021-067194","article-title":"Ketamine for the acute treatment of severe suicidal ideation: Double blind, randomised placebo controlled trial","volume":"376","author":"Abbar","year":"2022","journal-title":"BMJ (Clinical Res Ed)"},{"key":"B51","doi-asserted-by":"publisher","DOI":"10.1007\/s00108-025-02049-2","article-title":"Polypharmacy in patients with neuropsychiatric symptoms","volume":"67","author":"Schnieder","year":"2026","journal-title":"Innere Med (Heidelberg Germany)"},{"key":"B52","doi-asserted-by":"publisher","first-page":"12","DOI":"10.1001\/jamapsychiatry.2024.3241","article-title":"Slowing cognitive decline in major depressive disorder and mild cognitive impairment: A randomized clinical trial","volume":"82","author":"Rajji","year":"2025","journal-title":"JAMA Psychiatry"},{"key":"B53","doi-asserted-by":"publisher","first-page":"341","DOI":"10.1186\/s12888-024-05716-0","article-title":"Rapid and sustained antidepressant effects of intravenous ketamine in treatment-resistant major depressive disorder and suicidal ideation: A randomized clinical trial","volume":"24","author":"Zolghadriha","year":"2024","journal-title":"BMC Psychiatry"},{"key":"B54","doi-asserted-by":"publisher","DOI":"10.1159\/000548208","article-title":"Artificial intelligence-driven nephrology: The role of large language models in kidney care","volume":"57","author":"Zoccali","year":"2026","journal-title":"Am J Nephrol"},{"key":"B55","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0325982","article-title":"The sports nutrition knowledge of large language model (LLM) artificial intelligence (AI) chatbots: An assessment of accuracy, completeness, clarity, quality of evidence, and test-retest reliability","volume":"20","author":"Solomon","year":"2025","journal-title":"PloS One"},{"key":"B56","doi-asserted-by":"publisher","DOI":"10.3389\/fpsyt.2026.1817890","article-title":"Quality, reliability, and transparency of late-life depression videos on Chinese social media: a cross-sectional study of Douyin, Rednotes, and BiliBili","volume":"17","author":"Chen","year":"2026","journal-title":"Front Psychiatry"}],"container-title":["Frontiers in Psychiatry"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpsyt.2026.1956736\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,11]],"date-time":"2026-09-11T09:59:25Z","timestamp":1789120765000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fpsyt.2026.1956736\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,11]]},"references-count":56,"alternative-id":["10.3389\/fpsyt.2026.1956736"],"URL":"https:\/\/doi.org\/10.3389\/fpsyt.2026.1956736","relation":{},"ISSN":["1664-0640"],"issn-type":[{"value":"1664-0640","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,11]]},"article-number":"1956736"}}