Abbott and Google unveiled a partnership on August 11, 2026 integrating Lingo glucose sensor data with Google's AI Health Coach, targeting 115 million Americans with prediabetes and launching one of the largest real-world metabolic health studies to date.
An NIH-funded study used deep learning to analyze thrice-daily smartphone surveys from 62 patients in opioid addiction treatment, accurately forecasting relapse risk hours to days in advance, though researchers caution the small, non-diverse sample needs broader validation.
An AI tool embedded in Epic at NewYork-Presbyterian and Weill Cornell Medicine flags pregnant patients at risk for postpartum depression during routine prenatal visits, part of a wider 2026 research push toward predicting the condition before it starts.
Kintsugi, a startup that spent seven years and 30 million dollars building AI that detects depression from voice recordings, shut down in February 2026 after failing to win FDA clearance in time, offering a sobering lesson for AI psychiatric diagnostics.
Genomics England's Generation Study is sequencing newborns' full genomes and using AI to sort millions of genetic variants into actionable results, catching rare treatable conditions in about 1 in 200 babies as the NHS plans a decade-long national rollout.
CMS has launched WISeR, a six-state pilot using AI to screen traditional Medicare claims for medical necessity through 2031, and while officials say it targets wasteful spending, physicians and patient advocates worry it will replicate the high overturn rates and delayed care already seen with AI-driven Medicare Advantage denials.
A new Harvard-developed AI model called popEVE has flagged 123 previously unknown disease genes in a cohort of 30,000 undiagnosed patients, with 25 already confirmed by independent labs, offering a new path through the years-long diagnostic odyssey many rare disease families face.
Catholic Health in Buffalo is using RapidAI to alert stroke teams within minutes instead of the 45 minutes diagnosis used to take, though physicians insist the AI only prioritizes cases and never replaces standard diagnostic steps.
A JAMA study of 1,800 clinicians found AI scribes saved 16 minutes of documentation time per eight-hour day, far less than vendors have promised, with benefits varying sharply by specialty and consistency of use.
Lausanne University Hospital's AI system HERACLES re-screens every patient for sepsis every six hours. A study of nearly 100,000 patient stays found in-hospital mortality for flagged cases fell from 20.54% to 15.27%.