{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,27]],"date-time":"2026-10-27T07:48:54Z","timestamp":1793087334352,"version":"4.5.0"},"reference-count":33,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J Med Internet Res"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Deep research agents are autonomous large language model\u2013based systems capable of iterative web search, retrieval, and synthesis. They are increasingly positioned as the next major leap in medical artificial intelligence. In this viewpoint, we argue that while these agents mark progress in information access and workflow automation, they represent an incremental evolution rather than a paradigm shift. We review current applications of deep research agents in biomedical scenarios, including literature review generation, clinical evidence synthesis, guideline comparison, and patient education. Across these early use cases, the tools demonstrate the ability to rapidly gather and structure up-to-date information, often producing outputs that appear comprehensive and well-referenced. However, these strengths coexist with unresolved and clinically significant limitations. Citation fidelity remains inconsistent across models, with subtle misinterpretations or unreliable references still common. Their retrieval processes and evidence-ranking mechanisms remain opaque, raising concerns about reproducibility and hidden biases. Moreover, overreliance on artificial intelligence\u2013generated syntheses risks eroding clinicians\u2019 critical appraisal skills and may introduce automation bias at a time when medicine increasingly requires deeper scrutiny of information sources. Safety constraints are also less predictable within multistep research pipelines, increasing the risk of harmful or inappropriate outputs. Finally, current evidence is largely limited to proof-of-concept evaluations, with little evidence from real-life clinical deployment. We contend that deep research agents should be embraced as assistive research tools rather than pseudoexperts. Their value lies in accelerating information gathering, not replacing rigorous human judgment. Realizing their potential will require transparent retrieval architectures, robust benchmarking, and explicit educational integration to preserve clinicians\u2019 evaluative reasoning. Used judiciously, these systems could enrich medical research and practice; used uncritically, they risk amplifying errors at scale. We contend that deep research agents should be embraced as assistive research tools rather than pseudoexperts. Their value lies in accelerating information gathering, not replacing rigorous human judgment. Realizing their potential will require transparent retrieval architectures, robust benchmarking, and explicit educational integration to preserve clinicians\u2019 evaluative reasoning. Used judiciously, these systems could enrich medical research and practice; used uncritically, they risk amplifying errors at scale.<\/jats:p>","DOI":"10.2196\/88195","type":"journal-article","created":{"date-parts":[[2026,2,25]],"date-time":"2026-02-25T14:00:06Z","timestamp":1772028006000},"page":"e88195-e88195","source":"Crossref","is-referenced-by-count":2,"title":["Deep Research Agents: Major Breakthrough or Incremental Progress for Medical AI?"],"prefix":"10.2196","volume":"28","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-7633-4901","authenticated-orcid":false,"given":"Matthew Yu Heng","family":"Wong","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9300-573X","authenticated-orcid":false,"given":"Ariel Yuhan","family":"Ong","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4485-5580","authenticated-orcid":false,"given":"David A","family":"Merle","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9239-745X","authenticated-orcid":false,"given":"Pearse A","family":"Keane","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,3,26]]},"reference":[{"issue":"11","key":"R1","doi-asserted-by":"publisher","first-page":"2792","DOI":"10.7150\/ijms.111780","article-title":"Large language models in medicine: applications, challenges, and future directions","volume":"22","author":"Yu","journal-title":"Int J Med Sci"},{"issue":"8","key":"R2","doi-asserted-by":"publisher","first-page":"1930","DOI":"10.1038\/s41591-023-02448-8","article-title":"Large language models in medicine","volume":"29","author":"Thirunavukarasu","journal-title":"Nat Med"},{"issue":"12","key":"R3","doi-asserted-by":"publisher","first-page":"320","DOI":"10.3390\/bdcc9120320","article-title":"A systematic literature review of retrieval-augmented generation: techniques, metrics, and challenges","volume":"9","author":"Brown","journal-title":"Big Data Cogn Comput"},{"key":"R4","doi-asserted-by":"publisher","DOI":"10.1038\/d41586-025-00377-9","article-title":"OpenAI\u2019s \u201cDeep Research\u201d tool: is it useful for scientists?","author":"Jones","journal-title":"Nature New Biol"},{"key":"R5","doi-asserted-by":"publisher","DOI":"10.2196\/75666","article-title":"Exploring the dilemma of AI use in medical research and knowledge synthesis: a perspective on deep research tools","volume":"27","author":"Ong","journal-title":"J Med Internet Res"},{"key":"R6","unstructured":"Introducing deep research. OpenAI. URL: https:\/\/openai.com\/index\/introducing-deep-research\/ [Accessed 17-01-2026]"},{"key":"R7","unstructured":"Zhang W Li X Zhang Y et al. Deep research: a survey of autonomous research agents. arXiv. Preprint posted online on Aug 18, 2025. [doi: 10.48550\/arXiv.2508.12752]"},{"key":"R8","unstructured":"Huang Y Chen Y Zhang H et al. Deep research agents: a systematic examination and roadmap. arXiv. Preprint posted online on Jun 22, 2025. [doi: 10.48550\/arXiv.2506.18096]"},{"key":"R9","unstructured":"Gemini deep research. Gemini. URL: https:\/\/gemini.google\/overview\/deep-research\/ [Accessed 17-01-2026]"},{"issue":"11","key":"R10","doi-asserted-by":"publisher","first-page":"2250","DOI":"10.1111\/ans.70161","article-title":"OpenAI's 'Deep Research' for the generation of comprehensive referenced medical text: uses and cautions","volume":"95","author":"Jesudason","journal-title":"ANZ J Surg"},{"key":"R11","unstructured":"Wang Z Chen Z Yang Z et al. DeepEvidence: empowering biomedical discovery with deep knowledge graph research. arXiv. Preprint posted online on Dec 23, 2025. [doi: 10.48550\/arXiv.2601.11560]"},{"issue":"12","key":"R12","doi-asserted-by":"publisher","first-page":"7023","DOI":"10.1016\/j.acra.2025.08.014","article-title":"Evaluating large language models for radiology systematic review title and abstract screening","volume":"32","author":"Dogra","journal-title":"Acad Radiol"},{"issue":"1","key":"R13","doi-asserted-by":"publisher","first-page":"147","DOI":"10.1038\/s41433-025-04138-w","article-title":"Can \u201cDeep Research\u201d agents and general AI agentic systems autonomously perform systematic review and meta-analysis?","volume":"40","author":"Loke","journal-title":"Eye (Lond)"},{"key":"R14","doi-asserted-by":"publisher","DOI":"10.1111\/jdv.70035","article-title":"Assessment of Deep Research for dermatology literature reviews: deep concern over the hype","author":"Keplinger","journal-title":"J Eur Acad Dermatol Venereol"},{"issue":"4","key":"R15","doi-asserted-by":"publisher","first-page":"381","DOI":"10.1007\/s44411-025-00106-8","article-title":"Revisiting the role of review articles in the age of AI-agents: integrating AI-reasoning and AI-synthesis reshaping the future of scientific publishing","volume":"126","author":"Thurzo","journal-title":"Bratisl Med J"},{"key":"R16","unstructured":"Gao Y Xiong Y Gao X et al. Retrieval-augmented generation for large language models: a survey. arXiv. Preprint posted online on Dec 18, 2024. [doi: 10.48550\/arXiv.2312.10997]"},{"key":"R17","unstructured":"Yang R Wong MYH Li H et al. Retrieval-augmented generation in medicine: a scoping review of technical implementations, clinical applications, and ethical considerations. arXiv. Preprint posted online on Nov 8, 2025. [doi: 10.48550\/arXiv.2511.05901]"},{"issue":"8","key":"R18","doi-asserted-by":"publisher","first-page":"3025","DOI":"10.1002\/ksa.12711","article-title":"Evaluating DeepResearch and DeepThink in anterior cruciate ligament surgery patient education: ChatGPT-4o excels in comprehensiveness, DeepSeek R1 leads in clarity and readability of orthopaedic information","volume":"33","author":"G\u00fcltekin","journal-title":"Knee Surg Sports Traumatol Arthrosc"},{"issue":"11","key":"R19","doi-asserted-by":"publisher","DOI":"10.3390\/healthcare13111344","article-title":"Evaluating the reliability and quality of sarcoidosis-related information provided by AI chatbots","volume":"13","author":"Yetkin","journal-title":"Healthcare (Basel)"},{"key":"R20","unstructured":"Chen S Moreira P Xiao Y et al. MedBrowseComp: benchmarking medical deep research and computer use. Presented at: 39th Conference on Neural Information Processing Systems (NeurIPS 2025); Dec 6-7, 2025. 2025.URL: https:\/\/openreview.net\/pdf?id=pSDvq2ZgPk [Accessed 11-03-2026]"},{"key":"R21","unstructured":"Yu A Yao L Liu J et al. MedResearcher-R1: expert-level medical deep researcher via a knowledge-informed trajectory synthesis framework. Preprint posted online on Aug 20, 2025. [doi: 10.48550\/arXiv.2508.14880]"},{"key":"R22","unstructured":"Introducing IQVIA med-R1 deep research agent. IQVIA. URL: https:\/\/www.iqvia.com\/blogs\/2025\/12\/introducing-iqvia-med-r1-deep-research-agent [Accessed 18-01-2026]"},{"issue":"6","key":"R23","doi-asserted-by":"publisher","first-page":"415","DOI":"10.1097\/JCMA.0000000000001231","article-title":"AI\u2019s deep research revolution: transforming biomedical literature analysis","volume":"88","author":"Wang","journal-title":"J Chin Med Assoc"},{"key":"R24","doi-asserted-by":"publisher","DOI":"10.3389\/fpsyg.2025.1699320","article-title":"Cognitive offloading or cognitive overload? How AI alters the mental architecture of coping","volume":"16","author":"Chirayath","journal-title":"Front Psychol"},{"issue":"8","key":"R25","doi-asserted-by":"publisher","first-page":"786","DOI":"10.1056\/NEJMra2503232","article-title":"Educational strategies for clinical supervision of artificial intelligence use","volume":"393","author":"Abdulnour","journal-title":"N Engl J Med"},{"issue":"5","key":"R26","doi-asserted-by":"publisher","DOI":"10.70777\/si.v2i5.15569","article-title":"Superintelligent agents pose catastrophic risks: can scientist AI offer a safer path?","volume":"2","author":"Bengio","journal-title":"Superintell Robot Saf Align"},{"key":"R27","unstructured":"Chen S Li Z Han Z et al. Deep research brings deeper harm. arXiv. Preprint posted online on Oct 13, 2025. [doi: 10.48550\/arXiv.2510.11851]"},{"key":"R28","doi-asserted-by":"publisher","DOI":"10.1177\/21501319251332215","article-title":"The use of an artificial intelligence platform OpenEvidence to augment clinical decision-making for primary care physicians","volume":"16","author":"Hurt","journal-title":"J Prim Care Community Health"},{"key":"R29","doi-asserted-by":"crossref","unstructured":"Jagarapu J Babata K Chamarthi S Hoyt R . The accuracy and repeatability of OpenEvidence on complex medical subspecialty scenarios: a pilot study. medRxiv. Preprint posted online on Dec 4, 2025. [doi: 10.64898\/2025.11.29.25341091]","DOI":"10.64898\/2025.11.29.25341091"},{"issue":"5","key":"R30","doi-asserted-by":"publisher","first-page":"800","DOI":"10.1093\/eurjcn\/zvaf066","article-title":"Google\u2019s AI co-scientist and OpenAI\u2019s deep research: new partners in health research?","volume":"24","author":"Moons","journal-title":"Eur J Cardiovasc Nurs"},{"key":"R31","unstructured":"Du M Xu B Zhu C Wang X Mao Z . DeepResearch bench: a comprehensive benchmark for deep research agents. arXiv. Preprint posted online on Jun 13, 2025. [doi: 10.48550\/arXiv.2506.11763]"},{"key":"R32","unstructured":"Yao Y Wang Y Zhang Y et al. A rigorous benchmark with multidimensional evaluation for deep research agents: from answers to reports. arXiv. Preprint posted online on Oct 2, 2025. [doi: 10.48550\/arXiv.2510.02190]"},{"issue":"10","key":"R33","doi-asserted-by":"publisher","DOI":"10.25236\/AJCIS.2025.081008","article-title":"Applications of attention mechanisms in explainable machine learning","volume":"8","author":"Zhu","journal-title":"Acad J Comput Inf Sci"}],"container-title":["Journal of Medical Internet Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.jmir.org\/2026\/1\/e88195","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T17:14:59Z","timestamp":1782839699000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jmir.org\/2026\/1\/e88195"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,26]]},"references-count":33,"URL":"https:\/\/doi.org\/10.2196\/88195","relation":{},"ISSN":["1438-8871"],"issn-type":[{"value":"1438-8871","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,26]]},"article-number":"v28i11e88195"}}