India

Rocket Learning Foundation

About 35 million low-income children in India lack access to early education and are falling behind their higher-income peers in literacy and numeracy. Rocket Learning mobilizes digital communities of parents and teachers to support early childhood education. Using generative AI models and machine learning, Rocket Learning will develop an AI coach that can create localized academic content, automate grading, and offer personalized learning paths to increase education opportunities and improve learning outcomes for children across India.

Global

RAD-AID

Millions of people die every year due to lack of access to medical care. Doctor and health worker shortages in low- and middle-income countries mean that even when people do get radiology scans and X-rays, qualified professionals aren’t available to read them. RAD-AID will support low-resource hospitals to use AI to triage and interpret X-rays and scans and communicate test results, with a focus on respiratory disease and breast cancer. This grant will help RAD-AID improve diagnostic technology and the experience for more than 30 million patients in 20 countries.

Guatemala

Wuqu’ Kawoq | Maya Health Alliance

Guatemala has one of the highest neonatal mortality rates in Latin America, especially among the country’s indigenous Maya population. While many problems leading to fetal distress during delivery and poor fetal growth can be addressed when identified early, the technology needed to detect them isn’t readily available everywhere. Wuqu’ Kawoq and safe+natal are developing a toolkit that leverages machine learning to help midwives in rural settings detect complications in real time. The toolkit—a low-cost ultrasound and blood pressure monitor connected to a smartphone—will eliminate preventable deaths and complications by enabling mothers to get the help they need when they need it.

Australia

The University of Melbourne

A common challenge for patients with dementia is agitation when emotional and physical needs go unmet. Combining the therapeutic benefits of music with wearable sensors and AI, MATCH (Music Attuned Technology - Care via eHealth) is an adaptive, music-based tool that aims to decrease agitation and reduce medical costs. This grant will support development of the sensor technology and AI-enabled music adaptive system as well as a pilot study in Australia to prepare the project to scale globally.

India

Wadhwani AI

Over 6 million rural households in India depend on farming cotton crops, but every year, pests like pink bollworm destroy up to 30% of them. With prior support from Google.org, Wadhwani AI developed an AI-driven app to identify pests and recommend mitigation strategies; farmers’ profits increased by 20% and pesticide use decreased by 25%. This reinvestment will enable Wadhwani to roll out the technology across the country in partnership with the Ministry of Agriculture for 10 staple crops—mitigating the threat of hunger for billions of people by protecting critical crops.

Uganda

Makerere AI Lab

The shortage of lab technicians operating microscopes in low- and middle-income countries means that patients with illnesses like malaria, TB, and cancer often have limited access to quality diagnosis. As a result, some people don’t get the right treatment, putting their lives at risk, while others will get treatment they don’t need, leading to drug resistance. Makerere AI Lab will develop a 3D-printed adapter, compatible with any phone or microscope, that processes images using AI, making it easier to quickly diagnose health conditions and potentially reducing diagnosis time by 25%.

South Africa

IDinsight

While maternal mortality rates in South Africa have declined in recent years, new and expectant mothers still lack support for urgent pregnancy concerns. IDinsight with Reach Digital Health has built a natural language-powered question-answering service that reduces barriers to critical health information and prioritizes those most in need. With support from Google.org, IDInsight will reduce the number of mothers with urgent health issues, improve language accessibility, and create an open-source platform that helps other organizations provide underserved community members easy access to information.

Global

DHI A/S

Over 1 billion people depend on wetlands for a living; they’re home to about 40% of the world’s biodiversity; and they have tremendous capacity to slow climate change by absorbing and storing carbon. And yet they remain poorly mapped and characterized. Using open-source satellite imagery, other accessible data sources, and machine learning, DHI A/S, in collaboration with UNEP, will expand our knowledge of the world’s wetlands, working on the ground in five pilot countries and mapping wetlands on a global scale.

Kenya

EIDU GmbH

While in-person tutoring is out of reach for most children, particularly in low- and middle-income environments where resources are limited, digital learning tools have the potential to increase learning opportunities for every child even in areas with limited internet connectivity. Using low-cost smartphones and AI, EIDU GmbH will provide lesson plans and content for teachers and digital exercises, assessments, and personalized tutoring for up to 2 million pre-primary students and hundreds of thousands of primary school students in Kenya —all available off-line for hard-to-reach areas.

Vietnam

International Road Assessment Programme

Road traffic crashes are the leading cause of death among children and young people worldwide. The combination of poor road design and vehicle speeds puts children at risk on their daily journeys to school. Using AI and satellite imagery and street-view images to detect road safety risks, iRAP will provide a country-wide star-rating evaluation of road infrastructure around schools in Vietnam, with potential to scale to other countries. By making this data more accessible, they aim to inform new policies and investment in pedestrian-friendly roads that will minimize preventable harm.

Global

HURIDOCS

Common bodies of human rights information, such as legislation and policies, remain trapped in difficult-to-access stacks of paper, preventing human rights defenders from working efficiently and governments from making data-driven decisions. With support from Google.org, HURIDOCS developed Uwazi, a free and open source tool that provides access to this information at an unprecedented scale. With a reinvestment, HURIDOCS will continue integrating machine learning into Uwazi to raise awareness about patterns of violence and leverage important legal precedents across the globe.

Sub-Saharan Africa

Causal Foundry

Each year, nearly 300,000 women and 5 million newborns die of causes related to pregnancy and childbirth. Many of these deaths are preventable with low-cost healthcare and behavior changes. Causal Foundry will build a smartphone-based tool that uses machine learning to deliver personalized interventions and recommendations to help community health workers manage patient information and support behavioral changes to improve health outcomes. With this grant, they aim to improve care for 1.6 million existing patients in sub-Saharan Africa and reach 3 million new patients annually by 2026.

United Kingdom & United States

University of Surrey

Of the 7 million profoundly deaf people worldwide whose first language is sign language, 80% can’t properly comprehend the spoken languages in their country. The lack of digital accessibility and sign language translators makes it difficult to navigate everyday information and activities, including education, healthcare, employment, and transportation. Using generative AI, the University of Surrey and Signapse will automatically translate online and offline text into real-time, photo-realistic sign language videos, making quick translation affordable and easy to secure for the almost 600,000 Deaf people in the US and UK for whom Sign Language is their first language.

Global

International Rice Research Institute

Climate change poses a major risk to production of rice , the main staple food for over half the world’s population and main source of income for 144 million smallholder farmers. This project will use AI to identify ways to better utilize the world’s largest, most diverse collection of rice seeds and develop climate-resilient rice varieties that will better equip farmers to mitigate climate risks. The seeds will be made available, for free, to researchers and farmers. After 5 years from release, the projected economic benefit of adopting the new rice varieties in Asia and Africa is over $30 billion.

Sub-Saharan Africa

Jacaranda Health

Across Sub-Saharan Africa, new and expectant mothers lack access to clinically-accurate information to guide their care. Jacaranda Health’s SMS-based digital health service uses behavioral nudges and a natural language-powered helpdesk to triage questions for clinically urgent cases. The service currently triages up to 5,000 messages daily in Kiswahili and English and rapidly connects mothers with human agents. With this grant, they will use machine learning to refine the model, such as risk-prediction, and deploy it in 4 new African languages to support 3 million underserved mothers.