AI Ultrasound Is Reaching Mothers Who Never Had a Scan

AI guided ultrasound is a category of tools that guide non specialists through a pregnancy scan and automatically measure and interpret key indicators, and it matters because nearly 800 women die every day from preventable causes related to pregnancy and childbirth, with about 95 percent of those deaths occurring in low and lower middle income countries. Ultrasound during pregnancy identifies the complications that kill, including abnormal fetal position, twins, and problems with the placenta, so that a mother can deliver in a facility equipped to handle them. In much of the world, that scan never happens, because there is no sonographer for hundreds of kilometers. AI is changing who can perform this exam.

The most prominent effort is a program from Philips, which is developing AI powered ultrasound specifically to expand access to maternal health in underserved communities, with support announced in late 2023 from the Bill & Melinda Gates Foundation-backed initiatives funding the work.

The Sonographer Shortage

Performing and reading an obstetric ultrasound traditionally requires years of training. Sonographers must find standard imaging planes in a moving fetus, a skill of eye and hand that takes a long apprenticeship to develop. Low income countries train a fraction of the sonographers they need, and those who are trained overwhelmingly work in cities.

The consequence is that in parts of sub Saharan Africa and South Asia, most women go through entire pregnancies without any imaging. Complications that a twenty minute scan would reveal surface instead during labor, in a clinic with no blood bank or surgeon, which is precisely how maternal deaths happen.

Guiding the Probe in Kenya

The Philips program pairs a handheld ultrasound device with AI that does two jobs: it instructs the user in real time on how to move the probe to capture the correct standard views, and it automatically measures and grades the images, flagging findings that need escalation. The target users are midwives, nurses, and community health workers, not physicians.

The proving ground has been Kenya. During pilot trials there, Philips has reported that the technology supported better informed decision making for pregnant women in rural, underserved communities, with feedback from midwives pointing to increased confidence in conducting scans themselves. In 2025 the company described the pilot's positive results and its continued expansion of the program, alongside efforts to bring similar point of care ultrasound to underserved areas in the United States.

A Wider Movement

Philips is one player in a broader shift. Research published in the medical literature, including peer reviewed studies indexed by the National Library of Medicine, has repeatedly shown that nurses and midwives with brief training plus AI guided tools can acquire diagnostic quality obstetric images, and that automated measurements match those of sonologists. Handheld probe makers like Butterfly Network have run parallel programs training health workers across Africa, contributing to documented gains in maternal care in some deployments.

The research community has also tested AI interpretation directly. Studies have shown deep learning models estimating gestational age from ultrasound clips as accurately as trained sonographers, a finding with big implications: dating pregnancies accurately determines whether a preterm delivery is imminent and whether interventions like steroids for fetal lung maturity should be given.

Scan, Refer, Survive

The logic of the whole field can be stated in one chain: a midwife with a phone sized probe, an app that shows her where to aim it, an algorithm that reads the image, and a referral pathway for anything abnormal. Each link is now being built and tested in real health systems. None of it requires inventing new physics. It requires making an existing diagnostic exam usable by the health workers who are actually present when and where mothers need care.

Nearly every maternal death is preventable with timely care. The scan is often the first moment the risk becomes visible. For the millions of women whose pregnancies have never been imaged, AI carried in a midwife's bag may be what finally brings that first look within reach.