The system is intended to run on mobile devices with a handheld probe operated by people without specialist sonography training.
The study, published in JAMA Network Open, tested the model at Northwestern Medicine in Chicago and Jacaranda Health clinics in Nairobi. Researchers lightly tuned it using 120 Chicago participants, about 6% of the original training set, to adapt to Clarius C3 probe hardware and extend predictions beyond 36 weeks. No Nairobi data was used for tuning.
The evaluation covered 192 Chicago and 193 Nairobi pregnancies. Google reports a combined mean absolute error of 4.18 days, compared with 4.48 days for expert sonographer biometry; reported errors were 4.10 days in Chicago and 4.27 days in Nairobi. The source does not establish regulatory authorization, clinical deployment, improved maternal or fetal outcomes, or performance outside the studied settings and hardware.