Ranks distinct AI gain and problem claims from the published record. Scores reward impact, independent source strength, scale, confidence, and recency.
Researchers built SolenopsisDetector to automate identification of Solenopsis fire ants, which currently depends on taxonomic expertise. Using 8,300 images, they compared whole-body versus segment-based strategies, training YOLO detectors to localize ants and body parts and then classifying with ResNet, MobileNet and InceptionV3.
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.
On 2026-08-04, a peer-reviewed study reported development of a ferroptosis- and lipid metabolism-related prognostic signature for colorectal cancer using transcriptomic data from TCGA-COAD and GSE39582, WGCNA, and a machine learning pipeline of forward stepwise Cox regression combined with Random Survival Forest. The RiskScore was an independent predictor of overall survival and stratified immune features, and CRY2 emerged as a key signature gene validated by knockdown experiments.
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.
On 2026-07-02, a peer-reviewed paper in AI & SOCIETY introduced Value-Sensitive Citizen Science (VSCS), a framework that combines Value-Sensitive Design with citizen science to involve community members as co-researchers in AI development. It uses the Participatory Value-Cognition Taxonomy and extended scenario reasoning to translate local values into technical requirements and embeds governance for lifecycle oversight.
Published July 17, 2026, this perspective argues that quantitative prediction of biomolecular recognition requires moving beyond static structures to ensemble-based thermodynamic and kinetic observables. It reviews physics-based sampling under approximate Hamiltonians and modern machine learning models that learn from structural and bioactivity data.
On 2026-07-17, a review in Physical Chemistry Chemical Physics summarized machine learning force fields for inorganic crystalline materials, describing how they combine first-principles accuracy with classical force-field efficiency to enable atomic-level studies across structural prediction, physical properties, defects and interfaces, and phase transitions.
On 2026-04-03, a peer-reviewed article reported a systematic review and bibliometric analysis of 627 articles on human-AI decision-making. The authors identified two critical dimensions, AI-human dynamics and decision typologies, and proposed a novel conceptual framework comprising four paradigms: adaptive intuitive, programmed algorithmic, interpretive analytical, and integrative hybrid decision-making.
ace