Adversarial Fragility and Language Vulnerability in Clinical AI: A Systematic Audit of Diagnostic Collapse Under Imperceptible Perturbations and Cross-Lingual Drift in Low-Resource Healthcare Settings

• URL: http://arxiv.org/abs/2605.16993v1
• Date: Sat, 16 May 2026 13:33:47 GMT
• Title: Adversarial Fragility and Language Vulnerability in Clinical AI: A Systematic Audit of Diagnostic Collapse Under Imperceptible Perturbations and Cross-Lingual Drift in Low-Resource Healthcare Settings
• Abstract summary: This study presents the first systematic dual audit of two safety vulnerabilities in clinical AI: adversarial image fragility and cross-lingual diagnostic drift.<n>Using DenseNet121, fine-tuned on the COVID-QU-Ex chest X-ray dataset, we demonstrate that diagnostic accuracy collapses from 89.3% to 62.0%.<n>In a parallel language experiment, we tested Llama3.1:8b and NatLAS (NLASAT) on 20 COVID-19 clinical cases presented in Standard English, Nigerian Pidgin (Naija), and Yoruba-inflected English.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Current clinical artificial intelligence (AI) systems are evaluated almost exclusively on clean, standardised, English-language inputs, conditions that do not reflect the realities of healthcare delivery in low-resource settings. This study presents the first systematic dual audit of two orthogonal safety vulnerabilities in clinical AI: adversarial image fragility and cross-lingual diagnostic drift. Using DenseNet121, the architecture underlying CheXNet, fine-tuned on the COVID-QU-Ex chest X-ray dataset (85,318 images; COVID-19, Non-COVID Pneumonia, Normal), we demonstrate that diagnostic accuracy collapses from 89.3% to 62.0% under a Fast Gradient Method (FGM) perturbation of epsilon=0.021, a magnitude imperceptible to the human eye. Standard defensive strategies including Gaussian smoothing and ensemble voting failed to restore clinical safety. In a parallel language fragility experiment, we tested Llama3.1:8b and NatLAS (N-ATLAS) on 20 COVID-19 clinical cases presented in Standard English, Nigerian Pidgin (Naija), and Yoruba-inflected English. Both models exhibited significant accuracy degradation: Llama3.1:8b dropped from 80.0% to 65.0% on Pidgin; NatLAS, an African-context model, collapsed from 85.0% to 55.0%, with diagnosis consistency falling to 50%. These findings establish a quantitative failure envelope for clinical AI under conditions representative of Primary Health Centre (PHC) deployment in Nigeria, and motivate urgent calls for adversarially hardened, linguistically inclusive clinical AI architectures.

Related papers

• IyawoBench v2.0: Extended Diagnostic Evaluation of Large Language Model Clinical Triage in Nigerian Primary Care [0.0]
  We present IyawoBench v2.0, an extended diagnostic evaluation of large language model clinical triage.<n>We introduce a formal mathematical framework that decomposes triage safety into three distinct failure modes: Conservative Escalation Bias, Systematic Downgrade Bias, and Middle-Tier Instability.
  arXiv  Detail & Related papers  (2026-07-31T07:11:55Z)
• Aletheia: An Offline-First Clinical Decision Support System for Differential Diagnosis in Low-Resource Healthcare Settings [0.62914438169038]
  Aletheia is an offline-first clinical decision support system designed for low-resource healthcare contexts across sub-Saharan Africa.<n>It is built upon Qwen2.5-3B-Instruct, fine-tuned using Quantised Low- Rank Adaptation (QLoRA) on a curated dataset of 27,000 clinical reasoning samples.
  arXiv  Detail & Related papers  (2026-07-14T13:44:20Z)
• Clinical Validation of the Melanoscope AI Mobile Dermoscopy Clinical Decision Support System [0.0]
  Early detection of malignant skin lesions is critical for prognosis.<n>Dermatologist shortages in Russian regions limit screening coverage.<n>Mobile dermoscopy clinical decision support systems offer a promising approach.
  arXiv  Detail & Related papers  (2026-05-26T18:29:53Z)
• IyàwóBench: A Benchmark for Evaluating Large Language Model Clinical Triage Accuracy on Undifferentiated Febrile Illness in Nigerian Primary Health Settings [0.0]
  Undifferentiated febrile illness is the leading cause of primary care outpatient visits in Nigeria.<n>No validated benchmark exists for evaluating large language model (LLM) clinical triage reasoning.<n>We introduce IywBench v1.0, a dataset of 200 synthetic clinical vignettes across eight febrile illness categories.
  arXiv  Detail & Related papers  (2026-05-22T10:25:51Z)
• A Breast Vision Pathology Foundation Model for Real-world Clinical Utility [65.57568187389113]
  We present bfBRAVE, a breast-adaptive pathology foundation model developed and evaluated using a total resource of 101,638 breast wholeslide images.<n>BRAVE supported practical roles in the clinical workflow, including safe exclusion of low-risk cases from routine AI-assisted second-review rescue of initially missed positives and prioritization of cases for further assessment.
  arXiv  Detail & Related papers  (2026-05-06T07:44:39Z)
• Cerebra: A Multidisciplinary AI Board for Multimodal Dementia Characterization and Risk Assessment [56.62016795093786]
  Cerebra is an interactive multi-agent AI team that coordinates specialized agents for EHR, clinical notes, and medical imaging analysis.<n>It supports privacy-preserving deployment by operating on structured representations and remains robust when modalities are incomplete.<n>Cerebra significantly improved expert performance, increasing accuracy by 17.5 percentage points in prospective dementia risk estimation.
  arXiv  Detail & Related papers  (2026-03-23T05:46:45Z)
• Clinician input steers frontier AI models toward both accurate and harmful decisions [10.599240857217811]
  We evaluated 21 large language models (LLMs) across 8 frontier models on differential diagnosis generation and next step recommendations.<n>Expert context significantly improved correct final diagnosis inclusion across all 21 models.<n>In GPT-4o experiments, explicit clinician uncertainty signals improved diagnostic performance after adversarial context.
  arXiv  Detail & Related papers  (2026-03-14T23:47:53Z)
• Comparative Analysis of Deep Learning Architectures for Multi-Disease Classification of Single-Label Chest X-rays [1.1470070927586018]
  This study presents a comparative evaluation of seven deep learning architectures for multi-class chest disease classification.<n> ConvNeXt-Tiny achieved the highest performance (92.31% accuracy, 95.70% AUROC), while MobileNetV2 emerged as the most parameter-efficient model (3.5M parameters, 90.42% accuracy, 94.10% AUROC)<n>These findings demonstrate that high-accuracy multi-disease chest X-ray classification is achievable without excessive computational resources.
  arXiv  Detail & Related papers  (2026-03-11T07:52:36Z)
• A DeepSeek-Powered AI System for Automated Chest Radiograph Interpretation in Clinical Practice [83.11942224668127]
  Janus-Pro-CXR (1B) is a chest X-ray interpretation system based on DeepSeek Janus-Pro model.<n>Our system outperforms state-of-the-art X-ray report generation models in automated report generation.
  arXiv  Detail & Related papers  (2025-12-23T13:26:13Z)
• Evaluating Spoken Language as a Biomarker for Automated Screening of Cognitive Impairment [37.40606157690235]
  Alterations in speech and language can be early predictors of Alzheimer's disease and related dementias.<n>We evaluated machine learning techniques for ADRD screening and severity prediction from spoken language.<n>Risk stratification and linguistic feature importance analysis enhanced the interpretability and clinical utility of predictions.
  arXiv  Detail & Related papers  (2025-01-30T20:17:17Z)
• Utilizing Machine Learning Models to Predict Acute Kidney Injury in Septic Patients from MIMIC-III Database [0.0]
  Sepsis is a severe condition that causes the body to respond incorrectly to an infection.<n>For septic patients, approximately 50% develop acute kidney injury (AKI)<n>Models that can accurately predict AKI based on specific qualities of septic patients are crucial for early detection and intervention.
  arXiv  Detail & Related papers  (2024-12-04T22:05:35Z)
• Artificial Intelligence-Based Triaging of Cutaneous Melanocytic Lesions [0.8864540224289991]
  Pathologists are facing an increasing workload due to a growing volume of cases and the need for more comprehensive diagnoses. We developed an artificial intelligence (AI) model for triaging cutaneous melanocytic lesions based on whole slide images.
  arXiv  Detail & Related papers  (2024-10-14T13:49:04Z)
• Detection of subclinical atherosclerosis by image-based deep learning on chest x-ray [86.38767955626179]
  Deep-learning algorithm to predict coronary artery calcium (CAC) score was developed on 460 chest x-ray. The diagnostic accuracy of the AICAC model assessed by the area under the curve (AUC) was the primary outcome.
  arXiv  Detail & Related papers  (2024-03-27T16:56:14Z)
• Advancing COVID-19 Diagnosis with Privacy-Preserving Collaboration in Artificial Intelligence [79.038671794961]
  We launch the Unified CT-COVID AI Diagnostic Initiative (UCADI), where the AI model can be distributedly trained and independently executed at each host institution. Our study is based on 9,573 chest computed tomography scans (CTs) from 3,336 patients collected from 23 hospitals located in China and the UK.
  arXiv  Detail & Related papers  (2021-11-18T00:43:41Z)
• The Report on China-Spain Joint Clinical Testing for Rapid COVID-19 Risk Screening by Eye-region Manifestations [59.48245489413308]
  We developed and tested a COVID-19 rapid prescreening model using the eye-region images captured in China and Spain with cellphone cameras. The performance was measured using area under receiver-operating-characteristic curve (AUC), sensitivity, specificity, accuracy, and F1.
  arXiv  Detail & Related papers  (2021-09-18T02:28:01Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.