Google uses AI across multiple stages of medical research, from analyzing disease mechanisms at the molecular level to supporting diagnosis and public-health planning. The company says its teams have published 90 peer-reviewed papers over the past five years on understanding and treating diseases, garnering more than 90,000 citations.
The article, written by Yossi Matias, Google’s vice president and general manager of Google Research, and Pushmeet Kohli, vice president of science at Google DeepMind and chief scientist at Google Cloud, presents an overview of research tools and partnerships rather than announcing a single medical product.
From Genomes to Brain Maps
Google presents the AlphaGenome Atlas platform as an index of predictions about the effects of single-letter changes in the genome on molecular biology. Researchers from the University of Exeter used the platform to analyze data from more than 54,000 people and identify variants associated with levels of proteins including PLA2G7, which is linked to aging, and EGLN1, a cellular oxygen sensor.
In drug discovery, researchers at the Institute for Global Health in Barcelona used the AlphaFold model to virtually screen 30,000 compounds for activity against Chagas disease. The work helped identify two FDA-approved drugs for repurposing against the disease, which affects approximately 8 million people and can lead to serious complications if untreated.
Google also reconstructed part of a human brain at nanoscale resolution in collaboration with the Lichtman Lab at Harvard, and made the data available through Neuroglancer, along with maps of the brains of a fly, a larval zebrafish, and a songbird.
Shortening the Distance Between Research and Application
In a clinical trial conducted with Stanford University School of Medicine, the DeepVariant and DeepConsensus models helped achieve a genetic diagnosis using whole-genome sequencing in less than eight hours in time-critical cases. Google says the achievement also received recognition from Guinness World Records.
The company also presents AlphaProtein Novo, a machine-learning pathway for designing new enzymes from first principles. In laboratory tests, the system designed enzymes for synthesizing key components of medicines and others for breaking down toxic environmental plastic under harsh conditions.
Supporting Diagnosis and Public Health
In two studies published in Nature Cancer in collaboration with Imperial College London and the National Health Service, a mammography system was tested on 125,000 women. The system detected 25% of interval cancer cases that had historically been missed and showed the potential to reduce the image-reading burden on specialists by up to 40%. These are research findings and do not mean that the system has been adopted as a replacement for radiologists.
In cases of childhood leukemia, DeepSomatic confirmed known variants and discovered 10 previously undetected mutations when applied to real cases that had previously been sequenced at Children’s Mercy Hospital. Google also used geospatial models with environmental, weather, and mobility data to help predict disease outbreaks. In research from WHO AFRO and INRB in the Democratic Republic of the Congo, conversational geospatial models helped identify the risks of exposure to and transmission of Ebola more quickly.
Why Does This Trend Matter?
The actual change here is not merely the use of a language model in a clinic, but the expansion of AI’s role to cover drug discovery, interpretation of genetic variants, processing of diagnostic images, and allocation of public-health resources. However, the source presents the findings from Google’s perspective and does not explain in this article the details of regulatory approval or the extent to which each model can be generalized across different populations and health systems.
Google.org says it provided more than $60 million in 2026 to organizations using advanced models in disease research, including Makerere University and the University of Pennsylvania. The value of these initiatives remains tied to converting research findings into reliable clinical tools while maintaining independent validation and medical oversight.