Artificial intelligence
Artificial intelligence has been studied in 151 papers since 2019: 1 trial and 12 observational studies in people, 134 reviews; 115 in the past year. The newest paper, Nov 18, 2026: “Artificial Intelligence and Cardiovascular Disease Prevention: Newton's Second Law Readdressed with a Cautionary Note”.
Research on it is growing 57% faster than aging research as a whole: 446 papers in the latest year counted
16 new papers in 30 daysSee all 147 papers on artificial intelligence
At a glance
Reviews describe artificial intelligence predicting hip fractures, frailty, and dementia in humans, with independent labs reporting hip-fracture prediction. A human trial reports improved cancer detection and reduced screening workload; another shows no clear effect on gastric neoplasm detection. Authors include company employees, and papers declare company ties.
130 findings in 94 papers from 2019 to 2026, 39 of them since April 2026
Independent labs found the same change in people. 41 independent labs in 19 countries.
“AI did not improve detection rate of gastric neoplasm after pathological review (RR, 1.13; 0.92-1.38; 1.42 vs 1.25%; P = .25).”
None of its findings points the other way.
- 41 independent labs in 19 countries
- Funders: National Institutes of Health and others
- 5 of 60 findings with company ties
- Human trial1
- Humans12
- In silico3
Studies in people
1 study gave artificial intelligence to people and measured what changed.
| What changed | Design | People | Source |
|---|---|---|---|
no change: Has no effect on detection of gastric neoplasms down: Decreases blind spots2.52 to 1.07, P < .001 | Randomized trial | 29,514 peopleGeneral population; after pathological review | Gastroenterology, Mar 9, 2026 |
What did studies of artificial intelligence find?
77Studies of artificial intelligence measured 77 outcomes; the most studied are diagnostic accuracy (6 studies), breast cancer detection rate (3) and risk stratification (2).
Diagnostic accuracy
upin humans5upin reviews1
6 studies
Diagnostic accuracy
upin humans5upin reviews1
Artificial intelligence improves diagnostic accuracy in humans (Alzheimer's disease early diagnosis).
“Artificial intelligence (AI) and machine learning techniques may improve diagnostic accuracy by analyzing multiple biomarkers simultaneously”
ReviewAlzheimer's disease early diagnosis
Theoretical and Natural Science
Artificial intelligence improves diagnostic accuracy in breast cancer screening.
“the improved accuracy of diagnosing breast cancer”
Reviewin breast cancer screening
AI Driven Precision in Breast Cancer Screening: Transforming Early Detection
International Conference on Inventive Computation Technologies (ICICT)
Artificial intelligence improves diagnostic accuracy in humans (AUC 0.82 vs. 0.72).
“AI-driven 3D volumetric analysis outperformed traditional 2D DXA densitometry (AUC 0.82 vs. 0.72)”
Review
Journal of Education Health and Sport
Artificial intelligence improves diagnostic accuracy in humans (88-96% accuracy).
“AI tools in upper GI endoscopy achieve diagnostic accuracies of 88-96% for early esophageal and gastric neoplasia, outperforming non-expert endoscopists.”
Review
Clinical and experimental gastroenterology
Artificial intelligence increases diagnostic accuracy in humans.
“AI models, particularly DL applied to neuroimaging, electrophysiological data, speech and language features, biomarkers, and digital behavioral data, achieve high diagnostic accuracy in distinguishing MCI, Alzheimer’s disease, and healthy aging.”
Reviewapplied to neuroimaging, electrophysiological data, speech and language features, biomarkers, and digital behavioral data
Technologies
Artificial intelligence increases diagnostic accuracy (cardiovascular diagnostics).
“Key findings indicate that AI systems consistently achieve higher diagnostic accuracy, reduce inter-observer variability, and facilitate earlier detection of cardiovascular conditions, leading to improved patient outcomes.”
Reviewcardiovascular diagnostics
Annals of medicine and surgery (2012)
Breast cancer detection rate
upin humans3
3 studies
Breast cancer detection rate
upin humans3
Artificial intelligence is linked with better breast cancer detection rate in humans (screening digital breast tomosynthesis) (+1.34 (95% CI, 0.33 to 2.36)).
“CDR improved from 6.23 to 7.57, an increase of 1.34 (95% confidence interval [CI], 0.33 to 2.36; P < .01).”
Humansscreening digital breast tomosynthesis
Journal of the American College of Radiology : JACR
Artificial intelligence increases breast cancer detection rate in humans (mammography-based breast cancer screening).
“AI increased or maintained the breast cancer detection rate in several scenarios, including the identification of recurrences”
Reviewmammography-based breast cancer screeningbreast
Ibero Ciencias - Revista Científica y Académica
Artificial intelligence increases breast cancer detection rate in humans (AI-assisted digital breast tomosynthesis reading compared with human reading alone) (+3.8‰).
“DBT with AI-assisted reading compared to human reading alone increased the detection rate in a prospective study by +3.8‰”
ReviewAI-assisted digital breast tomosynthesis reading compared with human reading alonebreast
Digital breast tomosynthesis and reading with artificial intelligence
RoFo : Fortschritte auf dem Gebiete der Rontgenstrahlen und der Nuklearmedizin
Risk stratification
upin humans2
2 studies
Risk stratification
upin humans2
Artificial intelligence improves risk stratification in humans (compared with conventional clinical scores).
“machine-learning models integrating multimodality imaging data improve risk stratification over conventional clinical scores”
Reviewcompared with conventional clinical scoresmachine-learning models integrating multimodality imaging data
Artificial Intelligence in Coronary Artery Disease: From Detection to Precision Management
Methodist DeBakey cardiovascular journal
Artificial intelligence improves risk stratification in humans (detection).
“AI models integrate multimarker blood panels, routine laboratory indicators, ultrasound imaging, and cell-free DNA (cfDNA)-derived molecular signals to improve risk stratification beyond traditional biomarkers”
Reviewdetection
Frontiers in oncology
Screening workload
downin humans2−31% reduction
2 studies
Screening workload
downin humans2−31% reduction
Artificial intelligence decreases screening workload in humans (up to 31% reduction).
“The primary AI workflow could improve cancer detection by 10.4% (1 per 1,000), maintain the recall rate (0.8% reduction) and reduce workload by up to 31%.”
Human trialn = 10,889
Nature cancer
Artificial intelligence decreases screening workload in humans (screening mammography) (mean workload reduction of 42.9% (SD 1.5%)).
“There was a mean workload reduction of 42.9% (SD 1.5%) with AI integration.”
Reviewscreening mammography
Global Journal of Health Science
Recall rate
no changein humans2RR 1.19
2 studies
Recall rate
no changein humans2RR 1.19
Artificial intelligence was not clearly linked with recall rate in humans (screening digital breast tomosynthesis) (-0.01% (95% CI, -0.32% to 0.30%)).
“The recall rate was 6.97% pre-AI and 6.96% post-AI, a decrease of 0.01% (95% CI, -0.32% to 0.30%; P = .47).”
Humansscreening digital breast tomosynthesis
Journal of the American College of Radiology : JACR
Artificial intelligence showed no clear effect on recall rate in humans (screening mammography) (pooled RR 1.19, 95% CI 0.87-1.62, p = 0.275).
“The outcomes for recall rates were not significantly different overall with AI use (pooled RR 1.19, 95% CI 0.87-1.62, I2 = 98.6%, p = 0.275).”
Reviewscreening mammography
Global Journal of Health Science
Radiologist workload
downin humans2
2 studies
Radiologist workload
downin humans2
Artificial intelligence decreases radiologist workload in humans (AI-integrated workflows (triage and second-reader replacement)).
“Prospective evaluations of AI-integrated workflows, including triage and second-reader replacement, showed non-inferior cancer detection with significant workload reduction compared with standard double reading.”
ReviewAI-integrated workflows (triage and second-reader replacement)
European journal of radiology
Artificial intelligence decreases radiologist workload in humans (prospective trials in mammography).
“demonstrate non‐inferior cancer detection with large reductions in radiologist workload”
Reviewprospective trials in mammography
AI for screening in healthcare: promise and challenges
Abdominal radiology (New York)
Early identification of coronary artery disease
upin humans1
1 study
Early identification of coronary artery disease
upin humans1
Artificial intelligence improves early identification of coronary artery disease in humans (screening).
“In screening, deep-learning electrocardiography and electronic health record-based digital markers enable earlier identification of clinically significant disease.”
Reviewscreening
Artificial Intelligence in Coronary Artery Disease: From Detection to Precision Management
Methodist DeBakey cardiovascular journal
Mammography reading workload
downin humans1
1 study
Mammography reading workload
downin humans1
Artificial intelligence decreases mammography reading workload in humans (appropriately integrated into mammography screening).
“suggest that appropriately integrated AI can improve cancer detection and reduce reading workload without necessarily increasing false-positive outcomes.”
Reviewappropriately integrated into mammography screeningbreast
ARTIFICIAL INTELLIGENCE IN MAMMOGRAPHY: CURRENT APPLICATION AND FUTURE PERSPECTIVES
Acta Innovations
Cancer detection
upin humans1
1 study
Cancer detection
upin humans1
Artificial intelligence improves cancer detection in humans (appropriately integrated into mammography screening).
“suggest that appropriately integrated AI can improve cancer detection and reduce reading workload without necessarily increasing false-positive outcomes.”
Reviewappropriately integrated into mammography screeningbreast
ARTIFICIAL INTELLIGENCE IN MAMMOGRAPHY: CURRENT APPLICATION AND FUTURE PERSPECTIVES
Acta Innovations
Prediction of diabetes onset and diabetes-related complications
upin reviews1
1 study
Prediction of diabetes onset and diabetes-related complications
upin reviews1
Artificial intelligence improves prediction of diabetes onset and diabetes-related complications.
“AI-based models have demonstrated improved performance in predicting DM onset and diabetes-related complications by integrating electronic health records, imaging data, and longitudinal clinical variables.”
ReviewAI-based models
Artificial intelligence in diabetes: Clinical applications across the disease continuum
Journal of diabetes and its complications
Cancer detection rate
upin humans1
1 study
Cancer detection rate
upin humans1
Artificial intelligence improves cancer detection rate in humans (+10.4% (1 per 1,000)).
“The primary AI workflow could improve cancer detection by 10.4% (1 per 1,000), maintain the recall rate (0.8% reduction) and reduce workload by up to 31%.”
Human trialn = 10,889
Nature cancer
Blind spots
downin humans1
1 study
Blind spots
downin humans1
Artificial intelligence decreases blind spots in humans during upper endoscopy (2.52 to 1.07, P < .001).
“AI reduced blind spots number from 2.52 to 1.07 (P < .001)”
Human trialduring upper endoscopy
Gastroenterology
Detection of gastric neoplasms
no changein humans1
1 study
Detection of gastric neoplasms
no changein humans1
Artificial intelligence showed no clear effect on detection of gastric neoplasms in humans after pathological review.
“AI did not improve detection rate of gastric neoplasm after pathological review (RR, 1.13; 0.92-1.38; 1.42 vs 1.25%; P = .25).”
Human trialafter pathological review
Gastroenterology
All 77Fewer
What is artificial intelligence linked with?
37Studies link artificial intelligence with 37 measures, among them breast cancer, accuracy and atrial fibrillation.
Breast cancer
upin humans1no changein humans1
2 studies
Breast cancer
upin humans1no changein humans1
Artificial intelligence is associated with breast cancer in humans (2.9% vs 0.6%).
“malignancy rates were 0.6% (1/154) among non-flagged SVAs and 2.9% (2/68) among AI-flagged SVAs (P = .17).”
Humansbreastfemales
Canadian Association of Radiologists journal = Journal l'Association canadienne des radiologistes
Artificial intelligence is associated with breast cancer in randomised and paired prospective trials (rate ratio 1.13 (0.83-1.55)).
“Across randomised and paired prospective trials (k = 3), the pooled detection rate ratio was 1.13 (0.83-1.55)”
Reviewin randomised and paired prospective trials
Diagnostics (Basel, Switzerland)
Accuracy
upin humans2
2 studies
Accuracy
upin humans2
Artificial intelligence predicts accuracy in humans (AUC 0.890 (95% CI 0.858-0.915)).
“Pooled standalone AUC across 14 studies and 1,214,885 examinations was 0.890 (95% CI 0.858-0.915)”
Review
Diagnostics (Basel, Switzerland)
Artificial intelligence is associated with accuracy in detecting vitreoretinal diseases from retinal imaging systems (PEA 95.76% (95% CI 95.0-96.47), PESen 91.94%, PESpe 96.09%).
“yielding an overall PEA of 95.76% (95% CI: 95.0-96.47), PESen of 91.94% (95% CI: 90.72-93.08) and PESpe of 96.09% (95% CI: 95.27-96.79).”
Reviewin detecting vitreoretinal diseases from retinal imaging systems
Vitreoretinal disease detection using artificial intelligence: a systematic review and meta-analysis
International ophthalmology
Atrial fibrillation
upin humans2
2 studies
Atrial fibrillation
upin humans2
Artificial intelligence predicts atrial fibrillation in humans (handheld single-lead ECGs).
“AI applied to handheld single-lead ECGs can predict incident AF with accuracy similar to established clinical risk scores”
Reviewhandheld single-lead ECGs
Reimagining atrial fibrillation screening beyond age-based thresholds using AI
NPJ digital medicine
Artificial intelligence is associated with atrial fibrillation in humans (subclinical or paroxysmal AF).
“AI-enhanced electrocardiography (ECG) and wearable photoplethysmography devices allow early detection and long-term, non-invasive screening of AF”
Reviewsubclinical or paroxysmal AF
Indian pacing and electrophysiology journal
Frailty
mixedin humans1upin reviews1
2 studies
Frailty
mixedin humans1upin reviews1
Artificial intelligence predicts frailty in scoping review of 33 publications.
“Like the broader frailty literature, emerging AI tools lack a consistent definition of frailty, leading to design and implementation inconsistencies.”
Reviewin scoping review of 33 publications
Journal of the American Geriatrics Society
Artificial intelligence predicts frailty (moderate to high predictive performance).
“We found that AI models achieved moderate to high predictive performance in predicting frailty.”
Review
Advances of artificial intelligence in predicting frailty using real-world data: A scoping review
Ageing research reviews
Prostate cancer detection
upin humans1
1 study
Prostate cancer detection
upin humans1
Artificial intelligence predicts prostate cancer detection in humans (AUC 0.93–0.97).
“AI models achieved high diagnostic accuracy in breast (91%–98%), cervical (92%–96%), and prostate cancers (Area Under the Curve: 0.93–0.97).”
Review
Health science reports
Cervical cancer diagnostic accuracy
upin humans1
1 study
Cervical cancer diagnostic accuracy
upin humans1
Artificial intelligence predicts cervical cancer diagnostic accuracy in humans (Diagnostic accuracy 92%–96%).
“AI models achieved high diagnostic accuracy in breast (91%–98%), cervical (92%–96%), and prostate cancers (Area Under the Curve: 0.93–0.97).”
Review
Health science reports
Breast cancer diagnostic accuracy
upin humans1
1 study
Breast cancer diagnostic accuracy
upin humans1
Artificial intelligence predicts breast cancer diagnostic accuracy in humans (Diagnostic accuracy 91%–98%).
“AI models achieved high diagnostic accuracy in breast (91%–98%), cervical (92%–96%), and prostate cancers (Area Under the Curve: 0.93–0.97).”
Review
Health science reports
Downstream cardiovascular prevention bottlenecks
mixedin reviews1
1 study
Downstream cardiovascular prevention bottlenecks
mixedin reviews1
Artificial intelligence is associated with downstream cardiovascular prevention bottlenecks.
“this review highlights that AI mainly shifts the bottlenecks of prevention downstream, rather than removing them”
Review
Bottlenecks in cardiovascular prevention: Can artificial intelligence close the gap?
Archives of cardiovascular diseases
Recurrence risk
upin humans1
1 study
Recurrence risk
upin humans1
Artificial intelligence predicts recurrence risk in humans (treatment and prognosis).
“pathology-based, radiomics-based, omics-based, and multimodal AI models show potential for predicting homologous recombination deficiency (HRD), platinum response, recurrence risk, and survival outcomes”
Reviewtreatment and prognosis
Frontiers in oncology
Risk prediction
upin humans1
1 study
Risk prediction
upin humans1
Artificial intelligence predicts risk prediction in electronic health records.
“machine learning algorithms integrated into the electronic health records of the Unified Health System exhibit high precision in identifying patients at high risk of decompensation”
Reviewin electronic health records
Nursing Edição Brasileira
All 37Fewer
Funding
2 grants since 2019US Department of Veterans AffairsSwiss National Science FoundationCHF 904K
How this record changed
Summary rewritten as its findings changed.
As it read before
One in silico study reports increased quality-adjusted life-years with artificial intelligence in asymptomatic females over ten years. A review reports predictive accuracy in humans and a statistically inconclusive association with breast cancer.
Summary written from the findings.