Automating The Future

During the lockdown, Bateesa participated in a unique data collection project where the team listened to various radio stations, collecting information about the most commonly discussed topics. This data helped them integrate new words into their systems, further training their language processing AI.

One of his current projects involves building a translation pipeline. This system will feature various modules, including text translation, text-to-voice, voice-to-voice, and document summarization, all aimed at bridging communication gaps in local languages.

“The goal is to develop an AI system that can translate between different languages—whether text, voice, or video—allowing communication between languages like Acholi, Ateso, and English,” Tobius explains.

Although the tool is still under development and not yet available to the public, Tobius is optimistic about its potential. Once completed, he believes his AI tool will be invaluable for refugees, helping them communicate with locals more easily. He hopes to include additional languages from refugee communities in Uganda, further enhancing the tool’s reach.

More than Just AI

Bateesa isn’t the only one using AI to address real-world problems. A separate team of young innovators at Makerere AI Lab is applying AI to help farmers. Their project, Mvet, uses AI to detect livestock diseases in real-time by analyzing images and location data. This early disease detection allows farmers to isolate sick animals, preventing disease outbreaks and saving money.

“MVet is a project that is curating datasets on livestock health and testing. The datasets include geotagged images labeled with signs and symptom tags. By leveraging AI and real-time data, MVet helps veterinarians, researchers, and farmers monitor, diagnose, and respond to diseases in livestock. The goal is to improve livestock health outcomes and transform agriculture in Uganda.” says Dr. Daniel Mtenbesa the M-Vet project lead 

While crop farmers are benefiting from AI-powered tools for disease detection, livestock farmers in Africa still face significant challenges. The lack of digital data has hindered progress in livestock health innovations, leaving many farmers without reliable technological support.  

One major challenge limiting the adoption of digital technologies in agriculture for smallholder farmers in Africa is limited internet connectivity. Many rural areas lack the necessary infrastructure to support stable and affordable internet access, making it difficult for farmers to access digital extension services, online markets, and real-time agricultural information. Without reliable connectivity, the potential of digitalisation to improve productivity and market access remains largely unrealized, slowing progress in transforming agricultural practices.  

Livestock health is critical to farmers’ livelihoods, yet the lack of reliable data often makes monitoring and responding to diseases difficult.

MVet is working to address this gap by adapting its technology to feature phones, ensuring that farmers can access vital livestock health information even without internet access.  

“In the future, we plan to refine our AI so that if a farmer has a challenge, they can simply call in, speak with a voice bot, and receive instant feedback along with text messages explaining what is happening. This feature is still in development, but since we are currently in the testing phase, we hope to integrate it soon,” Dr. Daniel Mtenbesa explains.

Mvet has already been tested by over 1,500 farmers in Uganda, with a focus on providing rapid point-of-care diagnostics, image-based disease diagnosis, and disease surveillance tracking. 

The app is also breaking down barriers to veterinary care by providing farmers in remote areas with quicker access to veterinary doctors. This allows for faster disease identification and more effective treatment.

“The MVET project is really going to transform the lives of many farmers around the country. And like as we talk, we have the prototype already going on. I believe it’s going to be a life changer, what’s outside in the public. And right now, we require, like we request for everyone’s support. Like when, if this project is supported largely, it will really impact the future of this country.” Says Thomas Ogaw Dak who is a researcher and data scientist  with M-Vet.

In the testing phase, farmers have learned to navigate the Mvet app and can now digitize animal health records, eliminating the need for paper records. The app helps farmers track animal production, such as growth rates and milk yield, spotting potential issues early.

“We’ve developed AI that can analyze an image of a cow to determine its condition score, providing insights into its nutritional and health status. We’re now advancing this to estimate weight from images, helping farmers assess market readiness and track feeding effectiveness. While we’ve started with cattle, we aim to extend this technology to goats and other livestock in the future.” Say Arinda Cyrus a data scientist with M-Vet

Mvet is simplifying animal health monitoring and enabling better management, particularly in rural areas like Jinja and Masaka. Although the app is still being expanded across the country, the team is working on developing a sustainable model for farmers to purchase and maintain the technology.

The challenge 

Although spaces like the Makerere AI Lab are coming up with AI-powered technologies to tackle some of Uganda’s long-standing challenges, there is still a gap in adoption. 

One of the key challenges is accessibility, many of these innovations remain within research labs and are yet to reach the communities that need them most. Farmers, for instance, may not have the necessary digital literacy or infrastructure to utilize tools like M-Crop and M-Vet fully.   

The Ministry of Information, Communication and Technology notes that the internet penetration rate in Uganda stands at approximately 27%, with a total of 13.3 million internet users out of a population of about 48.66 million people in the country. Some communities in Uganda remain underserved, lacking access to affordable Internet, access devices, and technical knowledge to benefit from any available Internet and AI tools that require internet connectivity. 

Additionally, the cost of implementation and maintenance of AI-powered solutions remains a barrier. Scaling these technologies beyond pilot phases can be difficult without proper funding and policy support. 

“Bridging the gap between innovation and adoption will require collaboration between researchers, policymakers, and private sector players to make these solutions more accessible, affordable, and user-friendly.” says Dr. Daniel Mtenbesa, a research scientists with Makerere AI Lab 

Despite these challenges, the Makerere AI Lab continues to push boundaries, demonstrating the potential of AI in solving Uganda’s pressing issues. 

Makerere AI lab solving real world problem 

Bateesa credits Makerere AI Lab for nurturing his development in the field of AI. “It taught me, mentored me, and shaped me into the researcher I am today. Here, you’re surrounded by brilliant minds, including PhD students and experienced project leaders, which has been invaluable to my growth.”

The lab offers fellowships where experts share their work, providing invaluable guidance for both academic and practical pursuits in AI. Bateesa highlights the exposure he’s had to real-world problems, emphasizing the lab’s unique ability to not only develop models but also deploy them in the field to make a tangible difference.

“This lab serves as a pathway for young and senior scholars—whether at the undergraduate or graduate level—to research and experiment with AI and data science,”

“It is structured to achieve three key things: first, to identify a community of potential beneficiaries, often in health or agriculture; second, to pinpoint the real challenges they face; and third, to match these challenges with effective computational solutions. project lead and also a research scientist.” says Dr. Daniel Mtenbesa the M-Vet project lead

Mutenbesa’s research with the students focuses on resource allocation challenges in mobile environmental sensing. He also leads several AI projects related to agriculture, including livestock health monitoring, microfinance, and environmental data analysis. Through AI, the lab is enhancing livestock condition scoring, estimating animal weight, and expanding these innovations to include other animals beyond cattle.

The AI and Data Science Lab at Makerere specializes in applying AI to address critical challenges in the developing world, with a focus on areas such as disease detection, agricultural innovation, and environmental monitoring.

However, a significant challenge is that many of these technologies are not standalone solutions; they require development within a broader ecosystem that includes experts, end-users, and continuous learning.Without collaboration and thorough testing, innovations are at risk of failing to address real-world needs effectively. 

While the lab focuses on research and outreach, aligning AI-driven solutions with the specific needs of beneficiary communities presents a complex hurdle.

The lab’s ongoing efforts to test and refine these technologies, ensuring they are tailored to address the unique challenges in sectors such as health and agriculture, are essential to driving impactful, sustainable solutions.

Since its inception in 2009, the lab has evolved, collaborating with organizations like the National Agriculture Research Organisation (NARO) to tackle agriculture-related challenges using computer vision and machine learning. Their work like the M-Vet project has laid the foundation for further advancements in AI for agriculture.

As Bateesa and other young innovators continue to use AI to solve pressing challenges, the Makerere AI Lab remains a cornerstone of technological innovation, empowering the next generation of problem-solvers.