How AI is Powering a New Era of Community Health

In some of the world’s most underserved regions across India and Uganda, artificial intelligence (AI) is quietly transforming healthcare access. 

Two pioneering organizations in these regions, Visilant and VectorCam, are showing what the future of healthcare could look like. By equipping local health workers with powerful AI diagnostic tools, they are bridging the gap where doctors or specialist researchers are few, clinics are distant, and early care is often out of reach.

"There are a billion people living with vision loss that could be treated if they could get access to eyecare," said Jordan Shuff, the founder and Executive Director of Visilant, a social enterprise dedicated to scaling eye care access globally. "In India, there's one eye doctor for every 100,000 patients. If everyone were to get seen just once a year, that doctor would only have about four minutes per patient."

Visilant's innovation to solve this is simple yet powerful. “We have an attachment that goes on the smartphone that makes it really easy to take images of the eye,” Shuff said. “With it, we are able to capture three types of images — diffused light, a slit beam and blue light — to see different aspects of the eye.” 

The team is developing an AI model that can work alongside the tool to diagnose conditions and their severity. “Then the patient would either stay at home, or we would pull in an ophthalmologist or send the patient to a hospital,” Shuff said. 

Training community health workers to use the device takes less than a day. Formal training includes a half-day workshop covering eye diseases and basic exams. "At conferences, I hand the device to someone and they're using it in 30 seconds," Shuff said. 

The initiative started with Aravind Eye Hospital in 2020 during the COVID-19 pandemic, when traditional eye care services were suspended. Communities began taking blurry photos of their eyes on their phones and sending them to doctors via WhatsApp. "We saw the community's initiative but also the limitations," Shuff said. "So we built a better tool."

Now, Visilant works with multiple hospitals across India, 105 health workers are formally trained to use it, and the device has screened over 30,000 patients, saving 8 hours of travel time per patient.

"We meet weekly with all our hospital partners,” Shuff said.” They've been part of 30 or 40 iterations of the product. When they finally told us, 'No more changes. This is what we've been dreaming of,' that’s when we knew we were ready to scale.”

But the goal isn’t to replace eye doctors. "It's to save their time,” Shuff said. “Cataracts, which are the leading cause of blindness worldwide, can often be diagnosed without an ophthalmologist. Our AI can handle that triage, freeing up doctors to focus on complex cases."

Across the continent in Uganda, a different kind of health challenge buzzes in the air. VectorCam, a tech nonprofit born out of Johns Hopkins University, is tackling malaria by identifying its vector: the mosquito that carries the disease.

"Most investments in health systems in low- and middle-income countries are focused on protecting the human," said Sunny Patel, VectorCam's co-founder. "But we have determined there are a lot of cost-effective solutions when it comes to controlling diseases like malaria, dengue, and Zika that come not just from protecting the human, but from being able to control the mosquitoes in the environment.”

Traditional mosquito surveillance involves setting traps, identifying mosquitoes under a microscope and manually recording results.“The problem is that there are not enough experts who can identify malaria mosquitoes,” Patel said. 

With VectorCam’s 3D-printed smartphone attachment, health workers can now magnify and image mosquitoes with ease, allowing the phone to identify the species in seconds, Patel said. 

In a recent study in Uganda that has yet to be published, health workers identified mosquitoes in 18 seconds on average, Patel said. That's quicker than trained experts with microscopes, who typically need 30 seconds. 

"In many African countries, they only have a few different sentinel sites where they could study these mosquitoes, but then they use that to determine what the entire country’s policy should be. It can lead to missteps,” Patel said. “VectorCam makes it streamlined by bringing the community health workers, who are now abundant in rural Africa, and gives them the power to identify mosquitoes.”

The AI was trained on tens of thousands of mosquito images from Uganda and Mozambique. It runs directly on the phone, requiring no internet access. "Local health workers can get actionable data in real-time,” Patel said. 

The training to identify malaria mosquitoes using the tech is fast and practical. "Three days," Patel said. "Day one is about understanding malaria. Day two is learning to use the device. Day three is practice in the field. Some workers had never used smartphones before, but they picked it up quickly."

In their clinical trial, VectorCam captured 100 percent of the data needed for the findings to be useful within five days. In contrast, the traditional method collected only 55 percent of the necessary data after 14 days.

While their missions differ, Visilant and VectorCam share a core philosophy: empower local health workers with tools they can use to scale monitoring of systems and access to healthcare services. 

"We didn’t start with AI in mind," said Shuff from Visilant. "We started with a basic camera. But once we saw how well the first algorithms worked, we knew AI could help us scale."

Both organizations emphasize collaboration. Visilant co-developed its tool with India’s top eye hospitals. VectorCam works with Uganda's Ministry of Health and universities across Africa. Their tools are open source, locally trained, and easily integrated into national health systems.

For Patel, open source access to VectorCam tech is key for scaling success across borders. "We want a researcher in the Caribbean to be able to 3D-print our device, use it to upload mosquito photos, and get an AI model customized to their needs,” Patel said. “That’s the long-term vision.”

But both projects face systemic hurdles in terms of sustainability and scaling. Visilant is in the middle of validating its AI for wider deployment, and VectorCam is navigating pricing and scale in a sector dominated by donor funding.

"Everyone's excited," Patel said. "But who pays for it? Governments rely on external funding. We need a path that sustains itself."
Still, both organizations are optimistic. Visilant is preparing to launch manufacturing in India. "You can't have scale without India," said Shuff. "That's why we started here.”

VectorCam is beginning multi-district rollouts in Uganda and training students across Africa who are leaders in  entomology to expand adoption around the continent. In the push to make healthcare more equitable, these innovations demonstrate what AI can do when built with equity at the core.