In the bustling heart of Abuja, Nigeria, a quiet revolution is taking place within a modest office space. Here, Monsur Hussain and his dedicated team of software developers and AI researchers huddle together, their faces illuminated by the glow of laptop screens filled with complex code and machine-learning algorithms. This is a war room of sorts, but the battle being fought is not with weapons—it is against the rising tide of misinformation that threatens to overwhelm the digital landscape of West Africa. The team meticulously compares test results, refines algorithmic outputs, and debates the nuances of artificial intelligence, all with a singular goal: to teach machines how to distinguish fact from fiction in an increasingly complicated information ecosystem. Their work represents a crucial intersection of technology and journalism, where the future of truth in the digital age is being shaped one line of code at a time.
At the helm of this initiative is the Centre for Journalism Innovation and Development (CJID), a west African think-tank dedicated to enhancing journalistic standards through innovation and technology. Under Hussain’s leadership as head of innovation, the organization has developed Dubawa AI, a groundbreaking fact-checking tool launched in 2024 that promises to transform how journalists and the general public evaluate claims before they spread like wildfire across social media platforms. The challenge that initially seemed almost insurmountable was teaching AI to identify what constitutes a factual claim worthy of verification, as opposed to a simple question or an opinion. The team recognized early on that the distinction was crucial, particularly when dealing with claims that could have significant societal impact, such as those related to health, politics, and public policy. By training their models on years of verified fact-checks and developing sophisticated algorithms to detect patterns in previously reviewed allegations, they created a foundation upon which their entire system would later be built.
The significance of this work cannot be overstated when considering the scale of the challenge. With more than 148 million active internet subscriptions, Nigeria stands as Africa’s largest online market, making it a fertile ground for both information and disinformation. The advancement of generative AI technologies has made manipulated photos, videos, and audio increasingly convincing, placing unprecedented pressure on journalists to verify content before it reaches the public. For software engineer Japhet Johnson, the development of locally trained AI models is not just a technological achievement but a necessity born out of the limitations of global systems. These internationally developed models often fail to understand the nuances of African news presentation and may generate hallucinated or inaccurate information when dealing with regional contexts. The need for AI systems that comprehend how news is presented in Africa has become urgent, particularly as countries like Nigeria work to strengthen their data infrastructure to support these local innovations while ensuring compliance with national data protection regulations.
The practical application of Dubawa AI has taken the form of an accessible chatbot available on both the Dubawa website and the widely-used messaging platform WhatsApp, allowing users to verify claims within seconds. The system also includes an innovative audio tool that can extract and fact-check claims from recordings, making it invaluable for journalists working with interviews and speeches. Since its launch, the chatbot has successfully verified more than 6,000 claims and has been adopted across 32 African countries, with particularly strong growth in Ghana, Gambia, Liberia, Sierra Leone, and Senegal. The team’s approach to development was methodical, starting with a minimum viable product tested with journalists and fact-checkers who provided regular feedback that shaped subsequent iterations. This collaborative refinement process ensures that the tool remains practical and effective in real-world applications, rather than being purely theoretical in its approach.
For media researcher Ebube Glory Nwankwo, the real value of Dubawa AI lies in its transformative impact on daily work routines. Her responsibilities involve extensive media monitoring and gathering information from social media platforms—tasks that previously consumed days of manual effort. The ability to verify information using just her smartphone has revolutionized her workflow, eliminating the need to carry a laptop for fact-checking purposes. This accessibility represents a broader shift in how media professionals across the region can now approach their work, with the capacity for rapid verification becoming as mobile as the information they seek to validate. The tool’s effectiveness in saving time while maintaining editorial standards is precisely what Hussain envisions as the solution to reducing the spread of misinformation that currently flows freely through digital channels.
As with any AI system, the potential for generating false data posed significant concerns during development. Hussain and his team implemented robust guardrails and protective measures to ensure the model does not go out of context in various situations, a crucial step before making the tool public. However, for Hussain, the true measure of success goes beyond simple usage statistics or the number of claims verified. The deeper ambition is cultural transformation—building a society where people instinctively pause to verify information before forwarding it to others. This vision encompasses creating a form of accountability that monitoring and verifying claims can bring to public discourse. While the challenges remain considerable, from infrastructure limitations to the constant evolution of disinformation tactics, the work being done at CJID represents a significant step forward in the global fight against misinformation, one that could serve as a model for other regions facing similar challenges in the digital age.