Bias in medical AI: Implications for clinical decision-making
Open accessThe DOI matches its Crossref record; 124 of the 124 references it cites were found in a registry or an index; 30 of the 140 sources its claims cite were checked for support.
Abstract
Biases in medical artificial intelligence (AI) arise and compound throughout the AI lifecycle. These biases can have significant clinical consequences, especially in applications that involve clinical decision-making. Left unaddressed, biased medical AI can lead to substandard clinical decisions and the perpetuation and exacerbation of longstanding healthcare disparities. We discuss potential biases that can arise at different stages in the AI development pipeline and how they can affect AI algorithms and clinical decision-making. Bias can occur in data features and labels, model development and evaluation, deployment, and publication. Insufficient sample sizes for certain patient groups can result in suboptimal performance, algorithm underestimation, and clinically unmeaningful predictions. Missing patient findings can also produce biased model behavior, including capturable but nonrandomly missing data, such as diagnosis codes, and data that is not usually or not easily captured, such as social determinants of health. Expertly annotated labels used to train supervised learning models may reflect implicit cognitive biases or substandard care practices. Overreliance on performance metrics during model development may obscure bias and diminish a model’s clinical utility. When applied to data outside the training cohort, model performance can deteriorate from previous validation and can do so differentially across subgroups. How end users interact with deployed solutions can introduce bias. Finally, where models are developed and published, and by whom, impacts the trajectories and priorities of future medical AI development. Solutions to mitigate bias must be implemented with care, which include the collection of large and diverse data sets, statistical debiasing methods, thorough model evaluation, emphasis on model interpretability, and standardized bias reporting and transparency requirements. Prior to real-world implementation in clinical settings, rigorous validation through clinical trials is critical to demonstrate unbiased application. Addressing biases across model development stages is crucial for ensuring all patients benefit equitably from the future of medical AI.
Overview
CoChat Summary
TL;DR
This narrative review argues that bias in medical AI arises and compounds across the entire lifecycle—training data, model development and evaluation, deployment, and publication—and can worsen clinical decisions and health disparities. It catalogs mitigation strategies including diverse data collection, statistical debiasing, subgroup evaluation, interpretability, bias reporting, and clinical-trial validation.
Key findings
- Bias can enter at multiple stages: imbalanced sample sizes, nonrandomly missing data, uncaptured social determinants of health, biased labels, race/ethnicity in clinical algorithms, overreliance on whole-cohort performance metrics, publication biases, and implementation/end-user biases.
- The review cites specific examples and numbers: over half of published clinical AI models in 2019 used data from the US or China; radiology accounted for over 40% of AI publications in 2019; authors were 75% male; a MIMIC-III mortality model yielded recall rates as low as 25% under imbalanced racial representation; and the Epic Sepsis Model missed two-thirds of sepsis cases after deployment.
- Mitigation strategies discussed include large diverse data sets, oversampling methods such as SMOTE and ADASYN, imputation, SDoH capture, expert consensus labeling, subgroup analysis, bias-centered metrics like equalized odds and predictive parity, interpretability methods such as SHAP and LIME, adversarial debiasing, ongoing monitoring, and clinical trials for AI validation.
- Medical large language models pose unique bias risks because they are autoregressive, can reproduce biased or inaccurate training content, cannot assess their own outputs, and may generate inconsistent recommendations; statistical debiasing methods for clinical LLMs remain limited.
Methods
- Narrative review, not a systematic review, synthesizing literature and real-world or hypothetical examples across diverse medical domains and stages of the AI development pipeline.
- Includes Table 1 summarizing bias types by development stage, illustrative examples, potential clinical consequences, and mitigation strategies.
- No primary data collection, systematic search protocol, or quantitative meta-analysis is reported; the work draws on cited studies and examples.
Limitations
- Stated in the paper: more research is needed; there is a lack of standardized bias reporting guidelines; there are no explicit FDA regulations examining fairness or bias in medical AI outputs; and statistical debiasing methods for real-world clinical LLMs remain limited.
- Evident from the text: as a narrative review, it lacks systematic search and selection criteria and does not quantitatively synthesize evidence; examples are illustrative or hypothetical and not exhaustive; and the proposed mitigations are not empirically validated in this paper.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
17 Valid · 2 Not valid · 65 Not checked · 42 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Fewer studies have concentrated specifically on the implications and consequences of biased AI in healthcare [19–22].”
§ Bias in medical AI: Implications for clinical decision-making > Related work
What the sources say
- Not checked yetcites [19] RB Parikh, 2019, Addressing bias in artificial intelligence in health care, JAMA source
- Supportedcites [20] MD Abràmoff, 2023, Considerations for addressing bias in artificial intelligence for health equity, NPJ Digit Med source
The cited paper is an example of a study focused on biased AI in healthcare and its consequences. Its Introduction describes AI exacerbating health disparities and contributing to worse outcomes for Black patients; the “Bias in the healthcare process” section examines unequal outcomes across groups. The citation supports identifying this paper as one of the studies, rather than establishing a count of how many such studies exist. (LLM, conf=0.86)
- Partly supportedcites [21] R Daneshjou, 2021, Lack of transparency and potential bias in artificial intelligence data sets and algorithms: a scoping review, JAMA Dermatol sourceExperimental — accuracy not yet measured
The Discussion says that few dermatology AI studies report patient race, ethnicity, or skin tone, limiting assessment of potential bias. That supports a narrower gap in studying bias, but the paper does not establish that fewer studies across healthcare have examined the implications and consequences of biased AI. (LLM, conf=0.82)
- Not checked yetcites [22] AS Albahri, 2023, A systematic review of trustworthy and explainable artificial intelligence in healthcare: Assessment of quality, bias risk, and data fusion., Inf Fusion. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Standardized reporting checklists such as the Prediction Model Risk of Bias Assessment (PROBAST) tool have been developed to assess the risk of imbalanced data biases, which allows both developers and downstream users (e.g., physicians, hospitals) to better understand the suitability and limitations of AI-based systems in specific clinical settings [54–56].”
§ Bias in medical AI: Implications for clinical decision-making > Biases in training data > Imbalanced sample sizes
What the sources say
- Not checked yetcites [54] RF Wolff, 2019, PROBAST: a tool to assess the risk of bias and applicability of prediction model studies, Ann Intern Med source
- Partly supportedcites [55] T Hernandez-Boussard, 2020, MINIMAR (MINimum Information for Medical AI Reporting): Developing reporting standards for artificial intelligence in health care., J Am Med Inform Assoc sourceExperimental — accuracy not yet measured
The Introduction and General Principles of the MINIMAR Design say that reporting training-data characteristics can reveal potential biases and help assess whether an AI model generalizes to a clinical population. But the paper proposes MINIMAR; the passages found do not identify PROBAST or say that PROBAST was developed to assess biases from imbalanced data. No page numbers are available in our copy. (LLM, conf=0.83)
- Could not verifycites [56] Szczekocka E, Tarnec C, Pieczerak J, editors. Standardization on bias in Artificial Intelligence as industry support. 2022 IEEE International Conference on Big Data (Big Data); 2022: IEEE. sourceExperimental — accuracy not yet measured
The Standardization Specifics section says standards could help identify data-related biases, but it does not discuss PROBAST or a reporting checklist. Searches for PROBAST, imbalanced data, and clinical settings found no passage establishing the claimed purpose or use of that tool. (LLM, conf=0.76)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The researchers found that imbalanced representation of racial groups had a significant effect on model performance (yielding recall rates as low as 25%), raising critical concerns for the model’s potential application in real clinical settings [44].”
§ Bias in medical AI: Implications for clinical decision-making > Biases in training data > Imbalanced sample sizes
What the sources say
- Partly supportedcites [44] E Röösli, 2022, Peeking into a black box, the fairness and generalizability of a MIMIC-III benchmarking model., Sci Data. sourceExperimental — accuracy not yet measured
The Discussion reports significantly worse prediction performance for Black patients and says their low representation might partly explain the bias; it does not establish that imbalanced racial representation caused the performance gap. The Discussion also says the model identified only every fourth to fifth high-risk patient, but that low recall concerns the high-risk class overall, not a racial group. (LLM, conf=0.90)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“Furthermore, the degree of practitioner misclassification can vary along sociodemographic factors.”
§ Bias in medical AI: Implications for clinical decision-making > Biases in training data > Data labels and misclassification
What the sources say
- Supportedcites [84] GH Rauscher, 2013, Potentially missed detection with screening mammography: does the quality of radiologist’s interpretation vary by patient socioeconomic advantage/disadvantage?, Ann Epidemiol source
The INTRODUCTION states the goal was 'to examine whether variation in the quality of mammography interpretation as performed by the original reading radiologist varied by patient sociodemographic characteristics,' i.e. radiologist (practitioner) error by sociodemographic group. The RESULTS (p. 3, Table 3 discussion) report that potentially missed detection differed by minority ethnicity (+17 pp), income, education and lack of private insurance (+30 pp), and the DISCUSSION concludes the observed racial/ethnic and socioeconomic disparities 'shed light on differences in the quality of interpretation for different racial/ethnic and sociodemographic groups.' The Abstract's CONCLUSIONS affirm the association, so the hedged claim that practitioner misclassification 'can vary' along sociodemographic factors is supported; only the education association was attenuated in the selection-bias sensitivity analysis. (LLM, conf=0.90)
- Not checked yetcites [85] AM Kilbourne, 2005, Care without coverage: too little, too late., J Natl Med Assoc
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“There are several well-documented instances of insufficient evidence for race and ethnicity to be used as predictive factors in models.”
§ Bias in medical AI: Implications for clinical decision-making > Biases in training data > Race and ethnicity in clinical algorithms
What the sources say
- Partly supportedcites [92] MA Marzinke, 2022, Limited evidence for use of a black race modifier in eGFR calculations: a systematic review, Clin Chem sourceExperimental — accuracy not yet measured
The Discussion (page unavailable) concludes that there is little evidence for including a Black race modifier in eGFR calculations. This documents an instance of insufficient evidence for using race in a model, but the paper does not establish the broader claim of several instances involving race and ethnicity across models. (LLM, conf=0.85)
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
65 more claims · 8 with their cited sources checked
Create a free account to see all 126 claims and to verify the 110 references still open.
- It may state two separate claims as one. 41
- Probably not a substantial claim. 28
- The quote only partly states this claim. 1
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Generating educational resources without clinical context review risks reinforcing data biases [172]”
From Applications, Challenges, and Prospects of Generative Artificial Intelligence Empowering Medical Education: Scoping Review (2025) · cites this paper as [172]
Experimental — accuracy not yet measuredThe check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“In contrast, ensemble classification, combining multiple models, demonstrated improved accuracy and robustness [21,22].”
From Machine Learning in Maternity: Early Diagnosis of Placenta Accreta Spectrum (2026) · cites this paper as [22]
Experimental — accuracy not yet measured
Citations
Cites
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
SMOTE synthetic minority over-sampling technique
Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002) · Journal of Artificial Intelligence Research · doi:10.1613/jair.953 · Cited by 32,865 · cited 1×
“…which generates synthetic examples for the minority class by interpolating between existing minority class instances, and Adaptive Synthetic Sampling (ADASYN), which uses a similar method to SMOTE but focuses more on examples that are harder to classify (near the decision boundary) [49,50].”
MethodA Unified Approach to Interpreting Model Predictions
Lundberg, S., & Lee, S.-I. (2017) · arXiv · doi:10.48550/arxiv.1705.07874 · Cited by 23,707
Gauriat, C.-M., Pencolé, Y., Ribot, P., & Brouillet, G. (2024) · HAL (Le Centre pour la Communication Scientifique Directe) · doi:10.4230/oasics.dx.2024.27 · Cited by 12,586
Crenshaw, K. (2018) · Living with Contradictions · doi:10.4324/9780429499142-5 · Cited by 7,232
Dissecting racial bias in an algorithm used to manage the health of populations
Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019) · Science · doi:10.1126/science.aax2342 · Cited by 7,137 · cited 1×
“In reality, however, AI models are frequently found to be biased toward certain patient groups, leading to disparities not only in performance but also in potential or actual clinical benefits [7–9].”
BackgroundA Survey on Bias and Fairness in Machine Learning
Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2022) · ACM Computing Surveys · doi:10.1145/3457607 · Cited by 5,812 · cited 1×
“The topic of bias in AI has garnered significant attention in recent years, with multiple review papers addressing different aspects of this critical issue [14–18].”
BackgroundCrenshaw, K. (1998) · Feminism And Politics · doi:10.1093/oso/9780198782063.003.0016 · Cited by 5,672
Social determinants of health inequalities
Marmot, M. (2005) · The Lancet · doi:10.1016/s0140-6736(05)71146-6 · Cited by 5,086
ADASYN: Adaptive synthetic sampling approach for imbalanced learning
He, H., Bai, Y., Garcia, E. A., & Li, S. (2008) · 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence) · doi:10.1109/ijcnn.2008.4633969 · Cited by 4,521
Research works printed but not linked: 12
- Food, Administration D., 2019, Proposed regulatory framework for modifications to Artificial Intelligence/Machine Learning (AI/ML)-based Software as a Medical Device.
- D Roselli, 2019, editors. Managing bias in AI. Companion proceedings of the 2019, world wide web conference
- KPUL McFarling, 2024, Doctors use problematic race-based algorithms to guide care every day. Why are they so hard to change?, STAT
- Ribeiro MT, Singh S, Guestrin C. Model-agnostic interpretability of machine learning. arXiv preprint arXiv:160605386. 2016.
- Raji ID, Bender EM, Paullada A, Denton E, Hanna A. AI and the everything in the whole wide world benchmark. arXiv preprint arXiv:211115366. 2021.
- Chauhan VK, Clifton L, Salaün A, Lu HY, Branson K, Schwab P, et al. Sample Selection Bias in Machine Learning for Healthcare. arXiv preprint arXiv:240507841. 2024.
- Yang S, Wang Y, Chu X. A survey of deep learning techniques for neural machine translation. arXiv preprint arXiv:200207526. 2020.
- Topal MO, Bas A, van Heerden I. Exploring transformers in natural language generation: Gpt, bert, and xlnet. arXiv preprint arXiv:210208036. 2021.
- Schmidgall S, Harris C, Essien I, Olshvang D, Rahman T, Kim JW, et al. Addressing cognitive bias in medical language models. arXiv preprint arXiv:240208113. 2024.
- Ebrahimi S, Chen K, Asudeh A, Das G, Koudas N. AXOLOTL: Fairness through Assisted Self-Debiasing of Large Language Model Outputs. arXiv preprint arXiv:240300198. 2024.
- Li J, Tang Z, Liu X, Spirtes P, Zhang K, Leqi L, et al. Steering LLMs Towards Unbiased Responses: A Causality-Guided Debiasing Framework. arXiv preprint arXiv:240308743. 2024.
- Zhang Y-F, Yu W, Wen Q, Wang X, Zhang Z, Wang L, et al. Debiasing large visual language models. arXiv preprint arXiv:240305262. 2024.
Cited by
Items 201–300 of 716
Dietrich, N. T., Patel, D., Bellissimo, J., Loh, C. T., & Tyrrell, P. N. (2026) · Radiology Artificial Intelligence · doi:10.1148/ryai.250585 · Cited by 3
Perlmutter, J. W., & Joseph, M. (2025) · Education and human development · doi:10.5772/intechopen.1014025 · Cited by 3
Wang, R., Sun, X., Sun, Y., & Wang, J. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1794568 · Cited by 4
“Adjacent work in medical AI extends this warning by showing that bias may accumulate across the entire AI lifecycle, including feature construction, labeling, model development, evaluation, deployment, and publication practices (Cross et al., 2024).”
Dataset effects outweigh algorithmic effects in determining fairness of healthcare machine learning
Elgendi, M., Elkhalifa, A., Khandoker, A. H., Maruotti, A., Zhou, H., Menon, C., & Ward, R. (2026) · npj Digital Medicine · doi:10.1038/s41746-026-02723-1 · Cited by 6
“This reinforces that optimising for accuracy alone is insuf fi cient for ensuring equitable model behaviour 25,26,28 .”
Zhao, Y., Miao, Y., Luo, Y., Guo, R., Wang, H., & Wu, Y. N. (2026) · International Journal of Medical Informatics · doi:10.1016/j.ijmedinf.2026.106513 · Cited by 3
Lim, Y. X., Ming, L. C., Ng, N. C., Leow, S., & Ingle, R. G. (2026) · Discover Computing · doi:10.1007/s10791-026-10205-x · Cited by 3
“Models trained on biased or incomplete data may yield misleading predictions, reinforcing past failures rather than preventing future ones [14].”
Oropesa, A. L., Mendo-Lázaro, S., Gonzalez, A., del Barco, B. L., & Tapia, J. A. (2026) · Frontiers in Veterinary Science · doi:10.3389/fvets.2026.1816115 · Cited by 3
Wearable Flexible Sensors for Cardiovascular Disease Monitoring
Xie, X., Qu, X., Zhou, B., Zhang, B., Wang, Q., Bian, H., Shao, J., Cai, Y., & Dong, X. (2026) · Advanced Materials · doi:10.1002/adma.202512939 · Cited by 8
Zhao, Q., Jiang, W., Zhang, C., Ouyang, M., Wu, N., & Guo, Y. (2026) · Frontiers in Public Health · doi:10.3389/fpubh.2026.1863050 · Cited by 3
The digital therapist? LLMs and the future of clinical dialogue
Shi, E. R., Pei, R., & Barceló-Coblijn, L. (2026) · Language and Health · doi:10.1016/j.laheal.2026.100086 · Cited by 5
Takefuji, Y. (2025) · Coordination Chemistry Reviews · doi:10.1016/j.ccr.2025.216584 · Cited by 3
The Role of Artificial Intelligence in Surgery
Panossian, V. S., & Kaafarani, H. M. A. (2025) · Anesthesiology Clinics · doi:10.1016/j.anclin.2025.05.010 · Cited by 3
Elashram, R. E. M., Aldaghmi, O. A., & Mohammed, S. M. (2025) · Journal of Autism and Developmental Disorders · doi:10.1007/s10803-025-06947-0 · Cited by 4
Artificial Intelligence in Quality Improvement
Turek, G., & Sandberg, K. (2025) · Current problems in pediatric and adolescent health care · doi:10.1016/j.cppeds.2025.101756 · Cited by 4
Akkus Yildirim, B., Tutun, B., Durak, G., Yildirim, E. B., Uysal, E., Erturk, S. M., & Bagci, U. (2025) · Communications Medicine · doi:10.1038/s43856-025-01315-6 · Cited by 2
“However, it is crucial to acknowledge that LLMs are not without their own biases, primarily stemming from their training data, which can manifest in unexpected ways 21 .”
Gräßel, L., Yacob, S., Strantz, C., Prochaska, E., Debertshäuser, T., Rogge, A. A., Bäcker, S., Praßer, F., Unberath, P., Wehrle, J., et al. (2026) · npj Precision Oncology · doi:10.1038/s41698-026-01719-0 · Cited by 2
“If PRO data are used in AI-supported analytics or patient-similarity approaches, governance must also address algorithmic bias and representativeness 57 .”
Using ChatGPT in ethical dilemmas and policy-related complex decision making: Are we ready yet?
Tal, O., & Connelly, Y. (2025) · Health Policy and Technology · doi:10.1016/j.hlpt.2025.101041 · Cited by 3
Varela-Rodríguez, M., & Plage, S. (2025) · Journal of Cancer Survivorship · doi:10.1007/s11764-025-01843-z · Cited by 3
“In healthcare, generative systems have been shown to replicate demographic biases, raising concerns about the representational fairness of AI-generated medical imagery [53].”
Multi-objective swarm intelligence approach for bias mitigation in decision-making software
Vega-Cruz, L., & Vega-Rodríguez, M. A. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.129084 · Cited by 2
Gonzalez, L., & Dollarhide, L. (2025) · International Journal of Systemic Therapy · doi:10.1080/2692398x.2025.2549605 · Cited by 2
Shin, M., & Lee, D. (2025) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2025.2567966 · Cited by 4
“For example, unfair decisions may arise when medical AI systems allocate resources unevenly, when hiring algorithms disadvantage certain groups (Cross et al., 2024), or when AI-based hiring systems make discriminatory decisions—such as disadvantaging candidates based on gender, ethnicity, or…”
Chang, E., Xie, K., & Ellis, C. A. (2025) · Seminars in Neurology · doi:10.1055/a-2742-2430 · Cited by 2
Mendez, M., Jafarpisheh, N., DeMello, S., Lee, C., Dang, M., & Tyrrell, P. N. (2025) · PLoS ONE · doi:10.1371/journal.pone.0332716 · Cited by 4
Using artificial intelligence to simulate client interactions for psychology training
Hronis, A., Winiarski, N., Moustafa, A. A., & Kneebone, I. (2025) · Discover Psychology · doi:10.1007/s44202-025-00543-0 · Cited by 2
“AI systems trained on existing clinical data may inadvertently perpetuate biases and health-care disparities present in the training datasets, potentially leading to overrepresentation or underdiagnosis of certain conditions in specific demographic groups [33].”
A predictive model for cognitive decline using social determinants of health
He, Y., Leng, Y., Vranceanu, A., Ritchie, C. S., Blacker, D., & Das, S. (2026) · Journal of Aging Research and Lifestyle · doi:10.1016/j.jarlif.2025.100056 · Cited by 2
Quantum and Artificial Intelligence in Drugs and Pharmaceutics
Matarèse, B. F. E. (2026) · BioChem · doi:10.3390/biochem6010002 · Cited by 2
Audisio, A., ZHU, T. Y., Joeris, A., Sermon, A., IJpma, F. F. A., Giordano, V., Desai, D., Giannoudis, P. V., & Aprato, A. (2026) · Injury · doi:10.1016/j.injury.2026.113056 · Cited by 2
Sallam, M., Stanley, A., Snygg, J., Al-Shakerchi, H., Atragchi, O. A., Abusamra, R., & Sallam, M. (2026) · Recent Progress in Sciences · doi:10.70462/rps.2026.3.001 · Cited by 3
“These risks are amplified by a structural vulnerability since most genAI models are not intrinsically capable of distinguishing high-quality, evidence-based guidance from outdated, biased, or unverified sources [51, 54].”
Lazarus, J. V., Kopka, C. J., Nicolàs, A., Karim, S. A., Bansal, M. B., Betel, M., Dillon, J. F., Ginès, P., Krag, A. A., Miller, V., et al. (2026) · Nature Reviews Gastroenterology & Hepatology · doi:10.1038/s41575-026-01174-z · Cited by 2
Hubel, N. J., van der Heijden, T. G. W., Murauer, B., de Rooij, B. H., de Ligt, K. M., Verkooijen, H. M., Gernaat, S. A. M., Hoedjes, M., Arndt, V., van de Poll-Franse, L. V., et al. (2026) · Journal of Medical Internet Research · doi:10.2196/81424 · Cited by 3
Peterson, C., Anderson, C., Berdine, G. G., & NUGENT, K. M. (2026) · Journal of Electronic Resources in Medical Libraries · doi:10.1080/15424065.2026.2626575 · Cited by 2
“Furthermore, AI necessarily reflects the biases of the data on which it was trained (Cross et al., 2024).”
Cheema, M. N., Nazir, A., Harmanci, A. O., Harmanci, A. O., Cheema, Y., Masood, S., & Khokhar, F. A. (2026) · The Visual Computer · doi:10.1007/s00371-026-04385-2 · Cited by 2
“Studies emphasize the need for multilingual corpora and datasets inclusive of underrepresented populations [178].”
Pfahl, E., Pracha, N. S., Emlemdi, M., Le, P.-H. D., & Makary, M. S. (2026) · Biomedicines · doi:10.3390/biomedicines14030505 · Cited by 2
“Patients of a different race or ethnicity than that of the training population may have inaccurate results, which may not be appreciated until after clinical deployment of the model [86].”
Sidhu, R., Selvamogan, A., & Boddy, A. (2026) · Annals of The Royal College of Surgeons of England · doi:10.1308/rcsann.2026.0021 · Cited by 3
Salvaggio, G., Comelli, A., Albano, D., Galia, M., & Lalwani, N. (2026) · Abdominal Radiology · doi:10.1007/s00261-026-05457-6 · Cited by 2
Advances in Closed-Loop Artificial Intelligence for Healthcare
Das, D., Adams, S. D., Corva, D. M., Bucknall, T. K., & Kouzani, A. Z. (2026) · Electronics · doi:10.3390/electronics15071396 · Cited by 3
American Academy of Nursing Position Statement: Artificial Intelligence in Health Care
(2026) · Nursing Outlook · doi:10.1016/j.outlook.2026.102775 · Cited by 2
Are Natural Language Processing Tools Ready for Predicting Violence Toward Self or Others?
Grenier, S., St-André, M. F. A., Nguyen, B. T., Pierre, R., & Hudon, A. (2026) · Psychiatric Annals · doi:10.3928/00485713-20260324-03 · Cited by 2
Clinical Knowledge Representation in Data Science
Nguyen, T. A., Su, W., Rajagopalan, A. S., Abdurezak, N., Hewryk, O. S. I., & Romano, J. D. (2026) · Annual Review of Biomedical Data Science · doi:10.1146/annurev-biodatasci-092724-050810 · Cited by 2
Rabinowitz, A. R., Trauger, M., & Williams, M. (2026) · Archives of Physical Medicine and Rehabilitation · doi:10.1016/j.apmr.2026.04.033 · Cited by 2
Artificial Intelligence Applications in Pneumonia: Diagnosis and Outcome Prediction
Zhu, M., Bak, M., & Gao, C. A. (2026) · Current Pulmonology Reports · doi:10.1007/s13665-026-00411-9 · Cited by 2
“Moreover, bias in the data can significantly impact the model’s outcomes, particularly if certain populations (e.g., racial or ethnic minorities, rural populations, or under - represented age groups) are underrepresented in the train - ing datasets [ 67 ].”
Language Models in Veterinary Clinical Practice
Bollig, N., Lustgarten, J. L., & Venit, E. L. (2026) · Veterinary Clinics of North America Small Animal Practice · doi:10.1016/j.cvsm.2026.03.014 · Cited by 2
Gherghe, M. E., Grigore, A.-G., Timofticiuc, I.-A., Moise, A.-E., Andrei, C. A., Dragosloveanu, Ş., Nedelea, D.-G., Pulik, Ł., Anghel, C., Scheau, C., et al. (2026) · Bioengineering · doi:10.3390/bioengineering13060610 · Cited by 2
“Publication bias exists, favoring positive and technologically sophisticated results in AI research studies [102].”
Alignment of Policy, Practice, and Patient Safety for Trustworthy AI in Radiology
Doo, F. X., Davis, M., Poff, J. A., Lui, Y. W., Haines, K., & Towbin, A. J. (2026) · Radiology Artificial Intelligence · doi:10.1148/ryai.250982 · Cited by 2
Omeljanciuk, A., Peičius, E., Urbonienė, A., & Urbonas, G. (2026) · Healthcare · doi:10.3390/healthcare14121638 · Cited by 2
“In addition, an algorithm might work in one hospital but fail in another because every place uses the technology differently or simply because of the difference in the level of competence between AI users [67].”
Bridging the validation gap in artificial intelligence in radiology
Navarro-Ballester, A. (2026) · Exploration of Medicine · doi:10.37349/emed.2026.1001412 · Cited by 2
Natural Language Processing in Healthcare: From Unstructured Data to Clinical Intelligence
Kumar, D., Choudhary, N., Gupta, V., Tanwar, R., Bahmani, K., Antal, S., Jarora, S., Gupta, S., Shaikh, M. S., Webster, T. J., et al. (2026) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2026.200698 · Cited by 3
Agharia, S., Soroush, S., Ameen, D., & Zhou, Y. (2026) · Archives of Orthopaedic and Trauma Surgery · doi:10.1007/s00402-026-06400-6 · Cited by 2
Hajjo, R., Abusara, O. H., Sabbah, D. A., Zidan, R., & Bardaweel, S. K. (2026) · Journal of the American Pharmacists Association · doi:10.1016/j.japh.2026.103490 · Cited by 1
Naye, F., Ouellet, S., Yousefi, F., Cachinho, C., Bergeron, F., Bhattacharyya, O. K., Stamenova, V., & Sasseville, M. (2026) · Journal of Medical Internet Research · doi:10.2196/96882 · Cited by 2
ChatGPT and Large Language Models in Contemporary Nursing
Uchmanowicz, I., Aldossary, H. M., Lee, C. S., Figura, M., Jędrzejczyk, M., Maria, M. D., Pucciarelli, G., Molka, E., Straiton, N., & Kołtuniuk, A. (2026) · Journal of Clinical Nursing · doi:10.1111/jocn.70490 · Cited by 4
1,357 AI medical devices cleared, 3 actually tested on patient outcomes
Abulibdeh, R., Cajas, S. A., Celi, L. A., Gorijavolu, R., Izath, N., & Lunde, T. M. (2026) · PLOS Digital Health · doi:10.1371/journal.pdig.0001597 · Cited by 2
“Negative or inconclusive findings are seldom disseminated, reinforcing a literature dominated by algorithmic success [19].”
苗书英, Zhang, Z., & Bao, C. (2026) · Journal of Nursing Management · doi:10.1155/jonm/5288231 · Cited by 3
“These systems also pose inherent risks, including hallucinated or un-verifable content, prompt sensitivity, privacy leakage, automation bias, unequal performance across languages and cultural groups, and uncertainty about professional accountability [105–107].”
Okunev, M., Handelman, D., & Handelman, A. (2025) · Scientific Reports · doi:10.1038/s41598-025-15268-2 · Cited by 4
Explainable AI: Ethical Frameworks, Bias, and the Necessity for Benchmarks
Verhoeven, R., Bouisaghouane, W., & Hulscher, J. B. F. (2025) · European Journal of Pediatric Surgery · doi:10.1055/a-2702-1843 · Cited by 6
Are You Leading an Artificial Intelligence-Capable Healthcare Organization?
Arnaout, A. (2025) · Healthcare Management Forum · doi:10.1177/08404704251375388 · Cited by 7
“21 When AI is used it dictate clinical decision-making, it should be noted that over time, model accuracy can decline due to shifts in patient demographics, evolving clinical practices, or changing population health trends.”
A review on artificial intelligence in pharmaceutical sciences: opportunities and challenges
Kumar, V., Kumar, U., Kumar, R., Tiwari, N., & Gupta, J. K. (2026) · Drug Development and Industrial Pharmacy · doi:10.1080/03639045.2026.2670435 · Cited by 3
Iles, R. K., Zmuidinaite, R., Iles, J. K., & Nasser, S. (2025) · Preprints.org · doi:10.20944/preprints202508.1774.v1 · Cited by 2
Iles, R. K., Iles, J. K., & Nasser, S. (2025) · Preprints.org · doi:10.20944/preprints202508.2031.v1 · Cited by 2
Furlan, E., Zamperetti, N., Piccinni, M., Marra, A., Cotogni, P., & Giannini, A. (2026) · Journal of Anesthesia Analgesia and Critical Care · doi:10.1186/s44158-026-00395-7 · Cited by 2
Rosen, J. M. (2025) · Gastroenterology Clinics of North America · doi:10.1016/j.gtc.2025.01.001 · Cited by 2
Machine Learning for Predicting Medical Error Risks in Greek Surgery Departments
Michou, I., Maroulis, I., Hatzilygeroudis, I., & Koutsojannis, C. M. (2026) · Applied Sciences · doi:10.3390/app16115411 · Cited by 1
Mirza, H., Robson, A., Patel, A., Pagadala, V., Ledesma, M., Rodriguez, A., & Rao, S. (2025) · INNOVAPATH · doi:10.63501/sgexra86 · Cited by 2
Felici, A., Peduzzi, G., Pellungrini, R., & Campa, D. (2025) · European Journal of Cancer · doi:10.1016/j.ejca.2025.115716 · Cited by 1
Machine Learning in Medicine: A New Paradigm for Diagnosis and Treatment
Posunko, A., Chavarha, M., Bashkirova, L., Vasylyuk-Zaitseva, S., & Abramchuk, O. (2025) · Futurity Medicine. · doi:10.57125/fem.2025.09.30.01 · Cited by 1
“The problem of algorithmic bias is one of the most pressing challenges in modern digital medicine, as it can exacerbate existing inequalities in access to quality medical care [70].”
A Comprehensive Approach to Responsible AI Development and Deployment
Makhni, S., Rico, J., Cerrato, P. L., Hill, B., Overgaard, S. M., Wu, J., Fairbanks, J., Tripp, R., Redziniak, J., Blundo, R., et al. (2025) · Mayo Clinic Proceedings Digital Health · doi:10.1016/j.mcpdig.2025.100294 · Cited by 3
Perkin, P., & Köş, F. T. (2025) · Clinical and Translational Oncology · doi:10.1007/s12094-025-04090-x · Cited by 1
Eskandar, K. (2025) · Сибирский научный медицинский журнал · doi:10.18699/ssmj20250512 · Cited by 1
Translating Features to Findings: Deep Learning for Melanoma Subtype Prediction
Guermazi, D., Khemchandani, S., Wahood, S., Nguyen, C. V., & Saliba, E. (2025) · Dermatopathology · doi:10.3390/dermatopathology12040042 · Cited by 1
Lee, G., Cho, H.-H., Jeong, D. Y., Kim, J., Oh, Y. J., Park, S. G., & Lee, H. Y. (2025) · Journal of Thoracic Imaging · doi:10.1097/rti.0000000000000866 · Cited by 1
Healthcare Decision-Making Ecosystem Deengineering
Papalexandris, S., Fotiadis, T. A., Seremeti, L., Folinas, D., & Anastasiadou, S. (2025) · Cureus · doi:10.7759/cureus.97719 · Cited by 1
“This sociotechnical gap raises questions about accountability, transparency, and interpretability [95].”
Synthetic Data Augmentation for Exceptional Disease Using Generative Adversarial Networks
Gupta, S., Khanna, A., Mukhopadhyay, B., & Dubey, S. K. (2025) · IEEE International Conference on Contemporary Computing and Communications (InC4) · doi:10.1109/inc465408.2025.11256273 · Cited by 1
“To mitigate these, it is crucial to ensure diverse and representative synthetic datasets, implement statistical debiasing methods, and conduct thorough model evaluation across different subgroups [38].”
Navigating the new frontier: psychiatrist’s guide to using large language models in daily practice
Krishnadas, R., & Thilakan, V. (2025) · BJPsych Advances · doi:10.1192/bja.2025.10186 · Cited by 1
Natural language processing in healthcare communication
Gupta, R. K., Rajoriya, V., Vengurlekar, S., & Jain, S. K. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-43823-3.00005-4 · Cited by 1
Deceptive bias measurement in deep learning: Assessing shortcut reliance in TCGA cancer models
Kheiri, F., Rahnamayan, S., & Makrehchi, M. (2026) · PLOS Digital Health · doi:10.1371/journal.pdig.0001165 · Cited by 1
Building Trust in AI-Driven Organizations: A Strategic Approach to Ethical Management
Asrifan, A. (2025) · IGI Global eBooks · doi:10.4018/979-8-3373-3114-0.ch006 · Cited by 1
Adaptive Feature Representation Learning for Privacy-Fairness Joint Optimization
Ma, C., Dai, M., Guan, Z., Ye, Z., Hou, Y., Wang, X., & Huang, H. (2025) · Applied Sciences · doi:10.3390/app152413031 · Cited by 1
Beyond the Lab: Ethical, Legal and Social Implications in Health Biotechnology
Tabassum, A., Fatima, M., Fathima, S., Ameer, S., Fatima, Z., Fatima, N., Angelina, P., & Battula, A. (2025) · BIO Web of Conferences · doi:10.1051/bioconf/202520401003 · Cited by 1
AI in patient care: A narrative review of safety promises and pitfalls
Shopijen, E. (2025) · International Journal of Nursing Sciences · doi:10.1016/j.ijnss.2025.12.003 · Cited by 1
Alharmoodi, S., Nweran, M. A., El‐Kebbi, I. M., Everett, D., Ibrahim, S. M., Farhat, J., & Al‐Omari, B. (2025) · Patient Preference and Adherence · doi:10.2147/ppa.s567113 · Cited by 2
“75 Many algorithms are developed on small or single-center datasets, reducing their generalizability to diverse ethnic or socio-economic groups.”
Harto, B., & Ahida, R. (2025) · El-Rusyd Jurnal Sekolah Tinggi Ilmu Tarbiyah STIT Ahlussunnah Bukittinggi · doi:10.58485/elrusyd.v10i2.486 · Cited by 1
“The ultimate ideal is the development of fully objective artificial intelligence systems that do not reflect social biases or particular interests (Cross et al., 2024).”
李银娟 (2025) · Journal of Clinical Personalized Medicine · doi:10.12677/jcpm.2026.51048 · Cited by 1
Çiçek, V., Zhao, L., Tur, Y., Öz, A., Kılıç, Ş., Durak, G., Şaylık, F., Hayıroğlu, M. İ., Çınar, T., & Bağcı, U. (2026) · Medical Sciences · doi:10.3390/medsci14010033 · Cited by 1
“Inappropriate or unsafe AI-generated medical advice can lead to systemic consequences, especially when models are freely accessible to the public [7].”
Artificial Intelligence in Clinical Trials: From Protocol Design to Pharmacovigilance
Schulman, A., Fernández‐Torras, A., Gadiya, Y., Saeed, K., Mestres, J., & Tanoli, Z. (2026) · Springer eBooks · doi:10.1007/978-3-031-98022-0_24 · Cited by 2
To Use or Not to Use AI as Referees in Scientific Journals? Proposal for a More Efficient Model
Gozalbo, M., Soler, C., Onofre, N. S., & SORIANO, J. (2025) · INTELETICA · doi:10.4114/inteletica.vol2iss4pp1-11 · Cited by 1
“A hybrid model could shorten this timeline to 4-8 months, allowing more articles to be published annually, thereby enhancing journal competitiveness and financial [28].”
InfluentialIntegrating Scientific Realism and the Biopsychosocial Model in Biomarker Research for Depression
Garcia, G. (2026) · Issues in Mental Health Nursing · doi:10.1080/01612840.2025.2605643 · Cited by 1
“Although the model acknowledges biological contributors to depression, such as genetic predisposition, neurochemical dysregulation, and inflammatory signaling, its inherent breadth has made it difficult to translate into mechanistic hypotheses and empirical biomarker protocols (Cross et al., 2024).”
Educating Future Leaders in Drug Development about AI
Graff, M., Mollet, A., & Szucs, T. D. (2026) · Education and human development · doi:10.5772/intechopen.1014237 · Cited by 1
Xue, J., Wang, B., Wang, P., Niu, D., & Wei, B. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356416 · Cited by 1
“How inherited biases influence medical decision making remains an open question with critical implications for the real world [7].”
Ma, S., Zhou, Q., & Wu, H. (2026) · Advances in Medical Education and Practice · doi:10.2147/amep.s573537 · Cited by 2
“29,30”
Understanding Artificial Intelligence (AI) for the Electrophysiologist
Ramanathan, C., & Trayanova, N. A. (2026) · Indian Pacing and Electrophysiology Journal · doi:10.1016/j.ipej.2026.01.010 · Cited by 1
“Many models achieved high reported accuracy, sensitivity, or AUC during retrospective development and internal validation, yet showed substantial performance degradation during external testing or prospective evaluation [36].”
RISK-BASED TEST FRAMEWORK FOR LLM FEATURES IN REGULATED SOFTWARE
Zhou, Z. (2026) · doi:10.5121/csit.2026.160209 · Cited by 3
“This mirrors observations from AI validation literature, where expert review and clinical judgement are central to assessing performance and safety [3], [20], [21], [28], [30].”
InfluentialOyama, G., Tomizawa, Y., Tsunemi, T., Nojiri, S., Hatano, T., Sako, W., Hoshino, Y., Ueno, S.-I., Kamiyama, D., Oji, Y., et al. (2026) · Scientific Reports · doi:10.1038/s41598-025-33500-x · Cited by 1
“As highlighted in recent literature, such biases can arise from non-representative training data, underrepresentation of minority groups, or unclear model development processes, potentially leading to prejudiced outcomes or reinforcing existing disparities in patient care 21,22 .”
Terespolsky, S., Yassi, A., Ehrlich, R. I., Bruton, J., Lockhart, K., Wang, H., Klein, R., Sive, W., Statheros, J., & Spiegel, J. (2026) · Annals of Global Health · doi:10.5334/aogh.5064 · Cited by 2
A Comprehensive Review of Bias and Fairness in Artificial Intelligence for Oral Cancer Detection
Gadgimata, R. R., Jaiswal, S. J., & Nikita (2025) · International Conference on Electronics, Communication and Aerospace Technology (ICECA) · doi:10.1109/iceca66444.2025.11382669 · Cited by 1
“In order to make a fair assessment, the most prevalent one is the subgroup analysis, in which performance measures are contrasted with the performance measures of other demographic groups [9],[20].”
InfluentialBasile, G., Bolcato, V., Bambagiotti, G., Prevot, L. B., & Tronconi, L. (2026) · Bioengineering · doi:10.3390/bioengineering13020227 · Cited by 1
“And another scenario where competencies become generalized and there is an over-reliance on AI, which undermines the development of physician clinical skills, attentiveness, and risk assessment abilities [8,39].”
Neal, J. T., Hoffmann, R. M., Chiang, E. L., & Gravel, C. A. (2026) · Current Opinion in Pediatrics · doi:10.1097/mop.0000000000001555 · Cited by 1
Naveed, S., Husnain, M., Ikram, A., Ikram, S., Babar, M. J., & Alsubaie, N. M. (2026) · IEEE Access · doi:10.1109/access.2026.3672010 · Cited by 1
“Traditionally, such models rely heavily on structured data – such as diagnostic codes, laboratory results or demographic attributes [4].”
Artificial Intelligence in Military Graduate Medical Education: Trainee Perceptions and Current Use
Yabes, J., Lindholm, D. A., Ford, M. B., Kiley, J. L., & Geringer, J. L. (2026) · Military Medicine · doi:10.1093/milmed/usag099 · Cited by 1
Khan, S. H., Williams, M. S., & Lybarger, K. (2026) · PLOS Digital Health · doi:10.1371/journal.pdig.0001275 · Cited by 1
“In the health domain, such biases could influence how specific topics are represented or interpreted [58,59].”
Murray, B., Bauer, S. R., Gao, Y., Barreto, E. F., Henry, K., Blotske, K., & Sikora, A. N. (2026) · Frontiers in Medicine · doi:10.3389/fmed.2026.1787075 · Cited by 2
“Models that align poorly with clinical constructs may perform well but be difficult to apply in practice (4).”
Metadata
- Authors
- Cross, J. L., Choma, M. A., & Onofrey, J. A.
- Affiliations
- Yale University (US)
- Corresponding author
- Cross, James L.
- Venue
- PLOS Digital Health
- Publisher
- Public Library of Science (PLoS)
- Published
- 2024-11-07
- Volume
- 3
- Issue
- 11
- Pages
- e0000651
- ISSN
- 2767-3170 (ISSN-L)
- PMC
- PMC11542778
- OpenAlex
- W4404134492
- Cited by
- 607
- References
- 120
- Citations by year
- 2026: 382 · 2025: 222 · 2024: 3
- FWCI
- 13.83
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- PLOS Digital Health (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Keywords
- algorithmic bias, artificial intelligence in healthcare, clinical decision-making, health care disparities, biased training data, missing data, label bias, model evaluation, bias mitigation, model interpretability, social determinants of health, clinical validation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.