Inside a modest office in Abuja, Nigeria, the air is thick with concentration. Monsur Hussain and his team of software developers and AI researchers huddle around laptops, the glow of screens reflecting in their eyes. Lines of code scroll past, machine-learning models hum through test runs, and someone on the far side of the table presents a revised algorithm with a new set of results. These are not students working on a side project; they are the people behind Dubawa AI, a fact-checking tool designed to help journalists and ordinary citizens sort truth from falsehood in a region where information moves fast and trust can be hard to earn. Hussain is the innovation head at the Centre for Journalism Innovation and Development, a West African media think tank that uses investigative journalism, research and technology to improve the media landscape, promote accountability and combat misinformation across the region. He led the development of Dubawa AI, launched in 2024, and he describes his work as a constant conversation with the people who will actually use the tool. “Basically, deciding what to build, how to build it and what problem we are looking to solve, identifying the problem itself and working along with journalists and fact checkers to ensure that the solution we’re envisioning is the right one,” he says. It is not the kind of work that makes headlines, but it could determine whether the next viral rumour is caught in time. For Hussain, this isn’t about building technology for its own sake. It is about building trust in a region where misinformation can shape elections, influence health decisions and deepen social divides. The office may be modest, but the mission is anything but.
The name Dubawa comes from Hausa, a language widely spoken in northern Nigeria and across parts of West Africa. The word literally means “look,” but in the context of fact-checking it is best translated as “check” or “verify.” The name reflects the team’s core purpose. They began with a single, focused challenge: teaching artificial intelligence to detect factual claims that actually need verification. “There’s a difference between questions and claims,” Hussain explains. “What we want to focus on is claims.” The team trained their models to recognise statements that have societal consequences—health claims, political claims, and other assertions that could mislead people if left unchecked. This matters especially in Nigeria, where the Nigerian Communications Commission reports more than 148 million active internet subscriptions, making it the largest online market in Africa. At the same time, advances in generative AI have made altered photos, videos and audio recordings far more convincing. Journalists can no longer assume that a clip is real just because it looks real. For Japhet Johnson, a software engineer based in Abuja, locally developed AI models are essential because they are trained using regional information that many global models simply overlook. “Most of these AI models are trained maybe for Europe or the United States,” he says. “When we want to actually prove or get information, we find that most of those AI might hallucinate. They might not be able to give you good information about Africa. Regulation rules so that a lot of our data is stored internally.” In other words, building AI in Africa, with African data and African contexts, is not just a matter of pride; it is a matter of accuracy.
The technology behind Dubawa AI is designed to be practical and accessible. The team created a chatbot for the Dubawa website and for WhatsApp, allowing users to submit a claim and receive verification in seconds. They also built an audio tool that extracts claims from recordings, making it easier to fact-check speeches, interviews, radio broadcasts and viral voice notes. On the technical side, the platform was built using Python and connects directly to Dubawa’s database of previously fact-checked information. That database is a growing library of verified claims, corrections and context, and it gives the AI a solid foundation to draw on. For journalists, verifying a claim often means searching multiple databases, tracking down original sources, and cross-checking evidence before publication. It is slow, meticulous work, and under deadline pressure it can be overwhelming. Hussain believes that Dubawa AI can shorten that process while preserving the editorial judgment that human fact-checkers bring. “It’s about trying to reduce the misinformation that flies everywhere and trying to monitor what people are saying, what claims people are making and how we can put a control on it so that people understand there’s a form of accountability,” he says. The scale of the problem is clear: a study titled “Fake News and Public Trust in Online Media: A Study of Abuja, Nigeria” found that 75 percent of internet users polled frequently encountered fake news on their main social media platforms. In a country where smartphones are often the primary way people go online, having a fact-checking tool that works through WhatsApp is more than a convenience; it’s a necessity.
For the team’s own researchers, the impact has been immediate. Glory Nwankwo, a media researcher at the Centre for Journalism Innovation and Development, says Dubawa AI changed the way she monitors information gathered from social media and other online platforms. “My work as a researcher involves a lot of media monitoring and a lot of information on social media,” she says. “Doing it manually before Dubawa AI actually takes a lot of time, days sometimes.” The ability to verify information on a mobile phone has also transformed how she works when she’s away from her desk. “I don’t have to bring my laptop to check for information or sources. If I take my phone, I’m able to do it, and that really made my life easier.” This shift from days to seconds is about more than convenience. When false information spreads online, the window for correction is small. A rumour can be shared thousands of times before an official response is issued. A fact-checker who can verify on the go has a much better chance of catching a false claim before it becomes a headline. But the team is also aware that artificial intelligence is not a perfect solution. AI models can generate false data, and the same technology that powers fact-checking can also create deepfakes and manipulated content. That is why the team implemented protections before making the chatbot public. “We implemented certain changes and put in guardrails to ensure the model does not go out of context in a lot of situations,” Hussain says. The safeguards are not final; they are constantly tested and refined. For Hussain, however, the platform’s success depends less on the number of people using it than on whether it changes how people consume and share information.
Dubawa AI is part of a much larger mission. The Centre for Journalism Innovation and Development does investigative journalism, research and policy advocacy, and its technology team works closely with reporters and fact-checkers on the ground. This collaboration is what keeps the tool grounded. Journalists know what claims are circulating in their communities; researchers understand the patterns of misinformation; developers can build tools that respond to real needs. This is not a laboratory project isolated from everyday life. In West Africa, misinformation often spreads through oral channels: voice notes, radio broadcasts, sermons, community gatherings. The audio tool was designed with that reality in mind. The chatbot’s support for Hausa and other local languages helps bridge the digital divide that too often leaves women, older people and rural populations behind. As generative AI becomes more sophisticated, the boundary between real and fabricated will only get harder to draw. That makes local models like Dubawa AI more important, not less. Global AI systems are often trained on data from Europe and North America, leaving African contexts underrepresented. They may not understand local idioms, political history, cultural references, or the way a particular rumour mutates across borders. By building with local data and regional expertise, Hussain’s team aims to create a tool that understands the nuance of Nigerian society—and West Africa more broadly. The goal is not to replace human judgement but to support it, giving fact-checkers the speed and reach they need to keep up with the digital world.
At its core, Dubawa AI is not mainly about algorithms and code. It is about people. It is about a mother in Kaduna who receives a voice note with a fake cure for malaria and wonders whether to trust it. It is about a young voter in Lagos who sees a manipulated video of a politician and wants to know whether it is real. It is about a journalist in Abuja facing a deadline and needing to separate a claim from a lie. Monsur Hussain and his team are building a tool that treats information as something worth protecting. Their office may be modest, but their ambition is not. In a country with more than 148 million internet connections, the potential reach is enormous. Yet Hussain remains grounded. The success of Dubawa AI, he insists, will be measured by a change in culture. “I think what we want to build is a culture where people learn to check before they forward information, before they send something out.” That habit—pausing, checking, verifying—may be the most powerful antidote to misinformation. Technology can help, but it requires human choices. By making fact-checking faster, more local and more accessible, Dubawa AI is helping people make those choices. In doing so, it offers something that no algorithm can generate on its own: trust. And in a world where misinformation spreads at the speed of a tap, trust may be the most valuable currency of all.