When false information spreads online in English, artificial intelligence (AI) can often help identify it. But when the same misinformation appears in isiZulu, Sepedi or many other South African languages, the tools are limited.
A North-West University (NWU) PhD study by Dr Seani Ramanga is addressing that gap by developing multilingual AI that can detect misinformation across English, isiZulu and Sepedi, laying the groundwork for safer digital spaces in South Africa.
The research demonstrates that multilingual AI can identify misleading information across the three languages while highlighting the importance of improving machine translation for low-resource African languages. The study provides a framework that can be adapted to combat misinformation during elections, public emergencies and other events where false information can have serious consequences.
Dr Ramanga, a lecturer at the University of Pretoria, said the research was driven by the need to ensure that speakers of South Africa's indigenous languages are not left behind as AI becomes increasingly important in identifying misinformation.
"South Africa is one of the most linguistically diverse countries in the world, yet most AI systems for detecting misinformation are developed primarily for English. This leaves speakers of indigenous languages at a disadvantage when it comes to identifying false or misleading information online," she said.
Building trustworthy technologies
Her PhD research used misinformation surrounding the Covid-19 pandemic as a pilot case study to test the multilingual framework.
The pandemic illustrated how quickly false information can spread and the consequences it can have for people's lives, according to Dr Ramanga. While the initial research focused on health misinformation, the methods were designed to be applied across multiple sectors where misinformation presents a public risk.
The study found that multilingual AI can effectively detect misinformation in English, isiZulu and Sepedi while showing that the quality of machine translation plays a significant role in improving AI performance for low-resource languages.
"The framework provides a foundation for building more inclusive and trustworthy AI technologies for South African languages," Dr Ramanga said.
She said the findings could support government departments, fact-checking organisations, social media platforms and communities by enabling earlier detection of misleading information, improving access to credible information and helping people make informed decisions.
Standing proud on the world stage
The study has attracted international recognition. Dr Ramanga received the Google PhD Fellowship, one of the world's leading awards supporting doctoral research in AI. Her work also received the Best Poster Award at the 2025 Deep Learning IndabaX South Africa, earning her funding to present the research at the 2026 Deep Learning Indaba at Pan-Atlantic University in Lagos, Nigeria in August this year.
"Receiving the Google PhD Fellowship was a tremendous honour and a strong validation of the importance of conducting AI research for African languages. It reinforced my commitment to developing technologies that address challenges faced by multilingual and low-resource communities while also creating opportunities to collaborate with leading researchers around the world," she said.
Looking ahead, Dr Ramanga plans to expand the research to include more South African languages. Future work will focus on misinformation shared on social media during elections, public emergencies and other high-impact events. She also intends to extend the research to address emerging challenges such as AI-generated deepfakes, hate speech and other forms of harmful online content.
Beyond detection, Dr Ramanga aims to develop trustworthy multilingual AI systems, including large language models, retrieval-augmented generation systems and knowledge graphs that can support sectors such as healthcare, education, agriculture, governance and public services.
"My long-term vision is to ensure that African languages are fully represented in the next generation of AI technologies while contributing to safer and more reliable digital spaces for everyone."
Dr Seani Ramanga
