Ranks distinct AI gain and problem claims from the published record. Scores reward impact, independent source strength, scale, confidence, and recency.
A retrospective study in the Maryland Suicide Data Warehouse tested whether social determinants of health improve machine learning suicide prediction. Using 1214 suicide deaths and 815,544 living patients linked to census tract data, three algorithms were trained and validated across demographic, clinical, and individual and geographic SDoH inputs.
Researchers developed an unsupervised, interpretable AI framework to define tissue-agnostic cellular morphometric biomarkers that capture conserved tumor microenvironment organization across gastrointestinal organs. Discovered in colorectal cancer slides and validated in gastric and esophageal cancers in a 2,602-patient multi-center cohort, a 13-marker signature showed prognostic value and enabled risk stratification of precancerous lesions and early-stage cancers.
Researchers compared human-authored short stories with stories generated by GPT-3.5, GPT-4, and Llama 70b in response to the same prompts, using Burrows' Delta and clustering methods including hierarchical clustering and multidimensional scaling to visualize stylistic relationships.
In a quality improvement program evaluation, researchers used artificial intelligence to identify themes in interview data. By the publication date of 2026-08-24, the approach had produced four replicable themes grounded in the data, while also generating two consistently identified themes that were subtle misrepresentations.
A cross-sectional study tested four AI chatbots on 97 anonymized CBCT cases of jaw lesions, comparing performance on reconstructed 2D panoramic views and, for Manus, raw 3D DICOM data. Reports were scored for accuracy, relevance and feasibility, revealing statistically significant differences between systems.
In a prospective randomized crossover study at Carl R. Darnall Army Medical Center, 21 certified physician assistants interpreted 50 de-identified 12-lead ECGs with and without Queen of Hearts AI software by PMcardio. Diagnostic accuracy rose from 79.0% to 92.9% with AI, with sensitivity 95.4% versus 82.5% and specificity 90.5% versus 75.6%, and interrater agreement improved from kappa 0.58 to 0.86.
Published 4 August 2026 in Internal Medicine Journal, this peer-reviewed perspective examines rural mental health inequity in Australia and argues AI could help with earlier identification of distress and safer, more timely triage when used with telehealth and clinical decision support.
Published August 4 2026 in WORK, this peer-reviewed analysis examines AI integration into labour markets, migration governance and social protection systems. It argues AI functions as institutional infrastructure and shows how continuous classification, automated risk assessment and worker scoring can amplify structural inequalities when deployed at scale.
ace