MedGemma, an open-source medical AI model detailed in Nature Medicine, improved chest X-ray classification accuracy by 15.5 to 18.1 percentage points over previous models. An AI system applied to mammography identified 25% of cancers missed during routine screenings among 125,000 women while reducing specialist workload by 40%. A tool developed in partnership with Stanford University completed a genetic diagnosis in under eight hours, setting a Guinness World Record. Google Antigravity accelerates antimicrobial peptide identification as antibacterial resistance is projected to cause 10 million deaths annually by 2050. AlphaFold screening of 30,000 compounds contributed to the development of two FDA-approved drugs for Chagas disease.
Source: Read the original article
