AM-IA brings together news, research papers and applications relating to artificial intelligence published around the world into a single feed — and makes them available in your European language at the click of a button.
Scroll to browse · Click to filter
The additional payments for new technologies, paid two to three years after market entry, have never benefited as many AI tools as they did in 2026.
Effective adoption rose from 47% to 54% in less than a year, driven by the hope of recovering clinical time — but 71% of practitioners are still concerned about the accuracy of the answers.
Layoffs at Montefiore in the Bronx, protests at Kaiser Permanente: the largest health profession in terms of staff is getting organized in the face of the deployment of AI tools in hospitals.
Twelve major U.S. healthcare systems, covering nearly 20 million patients per year, are partnering to develop AI-assisted diagnostic workflows in response to the shortage of radiologists.
The medicine section of the Stanford report shows that a model with 111 million parameters outperforms the best previous methods, and that a genomic model with 200 million parameters outperforms a competitor with 40 billion.
The Danish laboratory wants to shorten the path between the identification of a therapeutic target and the first administration in humans using agentic systems.
A new, ready-to-use AI system published in Nature Cancer identifies eighteen different types of cancers with high accuracy from a reduced number of histological slides.
The meeting, set for Sept. 23, marks the first time an FDA panel has reviewed a multi-cancer early detection test.
In July 2026, Insilico Medicine launched a phase III trial involving 320 patients in 47 centers for the treatment of idiopathic pulmonary fibrosis based on its AI platform.
In the era of widespread clinical AI, the US agency is tightening the criteria under which a device can qualify for its fast-track procedure.
Published in Genome Medicine, the algorithm cross-references clinical notes and sequencing to classify candidate variants, including to reinterpret files that have remained undiagnosed.
DeepHealth's software (RadNet) obtained its 510(k) clearance in early August; the reading study reports +8% sensitivity and -37% interpretation time.
Out of 500 patient triage cases, the macro-F1 score of the unguided models peaks at 31-50%, and threshold errors persist despite instructions.
Published in Nature Medicine, a study on dermatological diagnostics reveals that the same explanatory tool can improve a clinician's performance and degrade a layman's judgment.
The randomized SAVE-O2 AI trial, conducted in 300 patients in four U.S. hospitals, shows that an automated oxygen flow adjustment holds the therapeutic range much better than caregivers.