医疗领域人机协同双刃剑效应及其影响因素———基于2010—2024年文献回顾分析

Double-edged sword effect and influencing factors of human-machine collaboration in healthcare: literature review from 2010 to 2024

摘要

目的 通过文献回顾明确医疗领域人机协同的内涵、特点及应用价值和关键影响因素,为优化临床实践提供科学支撑。方法 在数据库PubMed与Web of Science中,以人工智能(AI)、人机协同(“artificial-intelligence” OR “AI” OR “human-machine”)等为主题词,检索2010年1月至2024年10月期间相关文献,归纳医疗人机协同的内涵、积极效应、消极风险,并基于技术-组织-环境(TOE)理论阐述其影响因素。结果 共纳入93篇文献,内容涵盖AI应用场景、人机协同内涵、人机协同影响因素及作用等,明确了医疗领域人机协同的内涵是医生与AI系统通过动态协调与高效配合形成优势互补,进而实现医疗服务质量与效率提升的互动过程;对获取文献的分析结果显示,医疗人机协同具有提高诊疗效率、增强诊疗精准性、优化诊疗流程和推动医疗可及性共4项积极效应;同时存在责任归属模糊、医生技能退化、患者隐私泄露、加剧健康不公平、医患关系异化共5项潜在风险;并根据TOE理论分析共识别出15项关键影响因素,构建技术-组织-环境-个体四维影响因素分析框架。结论 明确了医疗人机协同的内涵,并揭示其“双刃剑”特征及影响因素分析框架,为医疗人机协同的实践应用及卫生政策制定提供参考。

Abstract

Objective To clarify the connotation, characteristics, application value and key influencing factors of human-machine collaboration in healthcare, so as to provide scientific support for optimizing clinical practice. Methods Taking artificial intelligence (AI) and human-machine collaboration (“artificial-intelligence” OR “AI”, OR “human-machine”) as the keywords, relevant literature published from January 2010 to October 2024 was retrieved from the databases of PubMed and Web of Science. The connotations, positive effects and potential risks of human-machine collaboration in healthcare were summarized, and its influencing factors were analyzed based on the TOE theory. Results A total of 93 articles were included, covering application scenarios of AI, the connotation of human-machine collaboration, as well as its influencing factors and functions. It was clarified that the connotation of human-machine collaboration in healthcare is an interactive process where clinicians and AI systems can achieve complementary advantages through dynamic coordination, thereby improving the quality and efficiency of medical services. The results showed that human-machine collaboration in healthcare have four positive effects, including improving diagnosis and treatment efficiency, enhancing diagnostic accuracy, optimizing clinical processes, and promoting healthcare accessibility. Meanwhile, five types of potential risks were identified, such as ambiguous responsibility attribution, clinical deskilling, patient privacy leakage, exacerbation of health inequity, and alienation of the doctor-patient relationship. Additionally, 15 key influencing factors were identified within a four-dimensional analysis framework from technology, organization, environment and individuals. Conclusion The review elaborates the connotation of medical human-machine collaboration, reveals its “double-edged sword” characteristics, and constructs a four-dimensional analysis framework of influencing factors, thus providing some reference for the practical application of medical human-machine collaboration and formulation of health policies.

关键词

医疗人工智能 / 人机协同 / 双刃剑效应 / TOE理论 / 影响因素

Key words

medical artificial intelligence / human-machine collaboration / double-edged sword effect / TOE theory / influencing factor

参考文献

[1] HAUG C J, DRAZEN J M. Artificial intelligence and machine learning in clinical medicine, 2023[J]. N Engl J Med, 2023, 388(13): 1201-1208.

[2] JOHANSSON J V, ENGSTRÖM E. Humans think outside the pixels' — radiologists' perceptions of using artificial intelligence for breast cancer detection in mammography screening in a clinical setting[J]. Health Informatics J, 2024, 30(3): 14604582241275020.

[3] ZHANG B, LAN X, WANG G, et al. A noise—suppressing neural network approach for upper limb human—machine interactive control based on sEMG signals[J]. Front Neurorobot, 2022, 16: 1047325.

[4] BULTEN W, BALKENHOL M, BELINGA J J A, et al. Artificial intelligence assistance significantly improves Gleason grading of prostate biopsies by pathologists[J]. Mod Pathol, 2021, 34(3): 660-671.

[5] STOWERS K, OGLESBY J, SONESH S, et al. A framework to guide the assessment of human—machine systems[J]. Hum Factors, 2017, 59(2): 172-188.

[6] AWA H O, OJIABO O U. A model of adoption determinants of ERP within T—O—E framework[J]. Inf Technol People, 2016, 29(4): 901-930.

[7] MASLEJ N, FATTORINI L, PERRAULT R, et al. Artificial intelligence index report 2024[DB/OL].arXiv(2024—05—29) [2025—05—23].https://doi.org/10.48550/arXiv.2405.19522.

[8] DIAW M D, PAPELIER S, DURAND—SALMON A, et al. A human—centered AI framework for efficient labelling of ECGs from drug safety trials[J]. IEEE Trans Biomed Eng, 2024, 71(5): 1697-1704.

[9] DUAN Y, CAI Y, PENG R, et al. Research on interaction and trust theory model for cockpit human—machine fusion intelligence[J]. Front Neurosci, 2024, 18: 1352736.

[10] DODEJA L, TAMBWEKAR P, HEDLUND—BOTTI E, et al. Towards the design of user—centric strategy recommendation systems for collaborative human—AI tasks[J]. Int J Hum Comput Stud, 2024, 184: 103216.

[11] ESMAEILZADEH P, MIRZAEI T, DHARANIKOTA S. Patients' perceptions toward human—artificial intelligence interaction in health care: experimental study[J]. J Med Internet Res, 2021, 23(11): e25856.

[12] KWON S, KIM J. Real—time upper limb motion estimation from surface electromyography and joint angular velocities using an artificial neural network for human—machine cooperation[J]. IEEE Trans Inf Technol Biomed, 2011, 15(4): 522-530.

[13] BLAAR M, JANß A, DELL'ANNA J, et al. Bottlenecks and needs in human—human and human—machine interaction — a view from and into the neurosurgical OR[J]. Biomed Tech (Berl), 2016, 61(2): 135-146.

[14] CALISTO F M, SANTIAGO C, NUNES N, et al. Breast—Screening—AI: evaluating medical intelligent agents for human—AI interactions[J]. Artif Intell Med, 2022, 127: 102285.

[15] ESMAEILI M, VETTUKATTIL R, BANITALEBI H, et al. Explainable artificial intelligence for human—machine interaction in brain tumor localization[J]. J Pers Med, 2021, 11(11): 1213.

[16] JIANG F L, JIA R Y, JIANG X J, et al. Human—machine interaction methods for minimally invasive surgical robotic arms[J]. Comput Intell Neurosci, 2022, 2022(1): 9434725.

[17] JIA X, MA Z, KONG D, et al. Novel human artificial intelligence hybrid framework pinpoints thyroid nodule malignancy and identifies overlooked second—order ultrasonographic features[J]. Cancers (Basel), 2022, 14(18): 4440.

[18] 陈浩, 陆清声 . 手术机器人的自动化研究进展[J]. 中国普通外科杂志, 2025, 34(1): 144-149.

[19] CHEN H, LU Q S. Progress in the automation of surgical robots[J]. Chin J Gen Surg, 2025, 34(1): 144-149.

[20] MARESCAUX J, DIANA M. Cybersurgery: human—machine integration for surgery of the future [La cyberchirurgie: l'intégration homme—machine pour la chirurgie du future][J]. Bull Acad Natl Med, 2013, 197(7): 1291-1301.

[21] Identity Theft Resource Center. ITRC data breach report 2023[R/OL]. (2024—01—24) [2025—05—23].https://github.com/OldLearner22/annual_security_reports/blob/main/Annual%20Security%20Reports/2023/ITRC—Annual—Data—Breach—Report—2023.pdf.

[22] 罗亦丹 . 奇安信: 2023年国内医疗行业泄露数据逾9亿条[N/OL]. (2024—05—14) [2025—05—29].https://www.bjnews.com.cn/detail/1715691646168519.html.

[23] LUO Y D. QI A X. Over 900 million data records leaked in China's medical industry in 2023[N/OL]. (2024—05—14)[2025—05—29].https://www.bjnews.com.cn/detail/1715691646168519.html.

[24] GAUBE S, SURESH H, RAUE M, et al. Do as AI say: susceptibility in deployment of clinical decision—aids[J]. NPJ Digit Med, 2021, 4(1): 31.

[25] BUDZYŃSKI K, ROMAŃCZYK M, KITALA D, et al. Endoscopist deskilling risk after exposure to artificial intelligence in colonoscopy: a multicentre, observational study[J]. Lancet Gastroenterol Hepatol, 2025, 10(10): 896-903.

[26] WANG S, NAN Y, ZHANG S, et al. Probing perfection: the relentless art of meddling for pulmonary airway segmentation from HRCT via a human—AI collaboration based active learning method[J]. Artif Intell Med, 2024, 154: 102930.

[27] SAUERBREI A, KERASIDOU A, LUCIVERO F, et al. The impact of artificial intelligence on the person—centred, doctor—patient relationship: some problems and solutions[J]. BMC Med Inform Decis Mak, 2023, 23(1): 73.

[28] RAINEY S, ERDEN Y J. Correcting the brain? The convergence of neuroscience, neurotechnology, psychiatry, and artificial intelligence[J]. Sci Eng Ethics, 2020, 26(5): 2439-2454.

[29] FERRARIO A, SEDLAKOVA J, TRACHSEL M. The role of humanization and robustness of large language models in conversational artificial intelligence for individuals with depression: a critical analysis[J]. JMIR Ment Health, 2024, 11: e56569.

[30] CABITZA F, CAMPAGNER A, SCONFIENZA L M. Studying human—AI collaboration protocols: the case of the Kasparov's law in radiological double reading[J]. Health Inf Sci Syst, 2021, 9(1): 8.

[31] STEYVERS M, KUMAR A. Three challenges for AI—assisted decision—making[J]. Perspect Psychol Sci, 2024, 19(5): 722-734.

[32] GÖNDÖCS D, DÖRFLER V. AI in medical diagnosis: AI prediction & human judgment[J]. Artif Intell Med, 2024, 149: 102769.

[33] 程乐 . 生成式人工智能的法律规制———以ChatGPT为视角[J]. 政法论丛, 2023(4): 69-80.

[34] CHENG L. Legal regulation of generative artificial intelligence: a ChatGPT perspective[J]. J Pol Sci Law, 2023(4): 69-80.

[35] CHA D, PAE C, LEE S A, et al. Differential biases and variabilities of deep learning—based artificial intelligence and human experts in clinical diagnosis: retrospective cohort and survey study[J]. JMIR Med Inform, 2021, 9(12): e33049.

[36] MAHMUD H, ISLAM A N, AHMED S I, et al. What influences algorithmic decision—making? A systematic literature review on algorithm aversion[J]. Technol Forecast Soc Change, 2022, 175: 121390.

[37] REZAZADE MEHRIZI M H, MOL F, PETER M, et al. The impact of AI suggestions on radiologists' decisions: a pilot study of explainability and attitudinal priming interventions in mammography examination[J]. Sci Rep, 2023, 13(1): 9230.

[38] ASAN O, BAYRAK A E, CHOUDHURY A. Artificial intelligence and human trust in healthcare: focus on clinicians[J]. J Med Internet Res, 2020, 22(6): e15154.

[39] INTROZZI L, ZONCA J, CABITZA F, et al. Enhancing human—AI collaboration: the case of colonoscopy[J]. Dig Liver Dis, 2024, 56(7): 1131-1139.

[40] MUDUMBAI S C, FANNING R, HOWARD S K, et al. Use of medical simulation to explore equipment failures and human—machine interactions in anesthesia machine pipeline supply crossover[J]. Anesth Analg, 2010, 110(5): 1292-1296.

[41] LAÏ M C, BRIAN M, MAMZER M F. Perceptions of artificial intelligence in healthcare: findings from a qualitative survey study among actors in France[J]. J Transl Med, 2020, 18(1): 14.

[42] BAN S, KIM Y, SEOMUN G. Digital health literacy: a concept analysis[J]. Digit Health, 2024, 10: 20552076241287894.

[43] 莫琳芳, 李喆, 甘辉亮, 等 . 全球视野下医疗人工智能中患者隐私和数据安全: 焦点与策略[J]. 海军军医大学学报, 2025, 46(8): 989-999.

[44] MO L F, LI Z, GAN H L, et al. Patient privacy and data security in medical artificial intelligence from a global perspective: focus and strategies[J]. Academic Journal of Naval Medical University, 2025, 46(8): 989-999.

[45] BIENEFELD N, KELLER E, GROTE G. Human—AI teaming in critical care: a comparative analysis of data scientists' and clinicians' perspectives on AI augmentation and automation[J]. J Med Internet Res, 2024, 26: e50130.

基金资助

国家卫生健康委医院管理研究所医疗人工智能临床应用研究项目(YLXX24AIF001)

中日友好医院高水平项目(2024-NHLHCRF-GL-12)