Google's ARDA retinal AI passes one million patients screened

Google's deep learning tool for diabetic retinopathy screening has crossed one million patients across three health systems: the Aravind Eye Care System in Madurai, India, Rajavithi Hospital in Bangkok, and Lions Outback Vision at the Lions Eye Institute in Western Australia, according to a Comment published in Nature Medicine on 23 September 2026.

The piece is lead-authored by Richa Tiwari of Google in Mountain View, with clinical co-authors from each deployment site: Kim Ramasamy and Ravindran Padmanabhan at the Aravind Medical Research Foundation, Paisan Ruamviboonsuk at Rajavithi Hospital, and Angus Turner at Lions Outback Vision. The authors describe "Scaling a successful deep learning tool across three highly distinct settings—in India, Thailand and Australia" as a source of "cross-cutting insights that may inform the expansion of healthcare artificial intelligence globally." Two researchers in our Who's Who shared the paper the day it dropped.

Rather than run the screening service itself, Google has licensed the model, Automated Retinal Disease Assessment or ARDA, to three regional commercial partners (Forus Health, AuroLab and Perceptra), who handle delivery on the ground. The stated target is "6 million AI-supported screenings to resource-constrained communities at no cost to patients" over the next decade.

The work is funded by Alphabet Inc.; several Google authors hold standard equity compensation, and Paisan Ruamviboonsuk discloses consulting work for pharmaceutical and biotech firms. The published abstract carries no per-site accuracy numbers, no country-level breakdown of the million screenings, and no fresh benchmark against ARDA's original clinical-trial performance; this is a retrospective lessons piece, not a diagnostic-performance paper.