In a modest Abuja office, Monsur Hussain and his team at the Centre for Journalism Innovation and Development (CJID) are redefining the fight against misinformation in West Africa. Recognizing that traditional journalism is struggling to keep pace with the rapid spread of digital falsehoods, the team developed Dubawa AI, a specialized tool launched in 2024. By leveraging machine learning models, the platform helps journalists and the general public evaluate the veracity of claims in real-time, moving beyond general AI to focus specifically on identifying and verifying socially impactful claims—ranging from health misinformation to political rhetoric—before they gain traction.
The urgency of this initiative is underscored by Nigeria’s status as Africa’s largest online market, with over 148 million active internet subscriptions. As generative artificial intelligence makes the creation of realistic deepfakes and manipulated content easier than ever, the pressure on newsrooms has reached a boiling point. Experts like software engineer Japhet Johnson argue that relying on global AI models—which are often trained on Western data and prone to “hallucinating” when applied to African contexts—is insufficient. Consequently, the CJID team has prioritized localized data infrastructure to ensure their tools accurately interpret the nuances of regional news while adhering to strict national data protection regulations.
At the core of this technological shift is a user-friendly chatbot accessible via the Dubawa website and WhatsApp. By integrating years of verified fact-checks from the CJID’s extensive database into a Python-based machine learning model, the developers have created a system that can detect trends in previously debunked allegations. The tool is designed to be highly responsive; users simply submit a claim, and the AI provides a verdict in seconds. This accessibility has already proven effective, with the tool verifying over 6,000 claims and gaining traction across more than 32 African countries, including significant adoption in Ghana, Gambia, Liberia, and Senegal.
The development process was rooted in a commitment to “human-in-the-loop” design. Hussain and his team intentionally developed a minimum viable product (MVP), refining it through rigorous testing and feedback from professional journalists and fact-checkers. This collaborative approach ensured that the technology serves as a practical aid rather than a replacement for editorial judgment. For researchers like Glory Ebube Nwankwo, the tool has been transformative, drastically reducing the time spent on manual media monitoring and allowing for instantaneous verification while away from a desk, thereby streamlining a formerly days-long labor process.
However, the team remains hyper-aware of the risks inherent in artificial intelligence. To mitigate the dangers of AI “hallucinations” and biased outputs, the developers have implemented strict guardrails to keep the model within verified contexts. These technical protections are a response to the dire reality of the local digital landscape, where studies have shown that 75 percent of internet users in Abuja frequently encounter fake news on social media. By automating the preliminary legwork of tracking down original sources and cross-referencing databases, the AI allows journalists to focus on high-level investigative tasks and accountability journalism.
Ultimately, Hussain views the success of Dubawa AI not merely as a metrics-based achievement, but as a catalyst for a broader societal shift. The goal is to instill a culture of skepticism and verification, where individuals pause to “check” before they “share.” By lowering the barrier to entry for fact-checking through mobile technology, the CJID hopes to empower citizens to act as their own gatekeepers. As the platform continues to iterate and expand, it stands as a critical defense mechanism in Africa’s fight for information integrity, proving that local solutions are essential for addressing the continent’s unique digital challenges.