The GSMA and Zindi, an AI challenge platform focused on emerging markets, have launched a competition aimed at identifying vulnerabilities in large language models when deployed across Africa’s more than 2,000 languages — a gap organizers say poses a significant blind spot in global AI safety standards.
The African Trust and Safety LLM Challenge was announced March 4 at MWC Barcelona. Running through April 19 on the Zindi platform, the competition invites data scientists to generate structured adversarial prompts and safety classifications that stress-test AI models across underrepresented African languages and code-switched contexts — where speakers blend multiple languages in a single conversation. The results will feed into a reusable, Africa-focused AI trust and safety benchmark. A total prize pool of $5,000 is on offer, with Zindi’s community of more than 100,000 data scientists and AI practitioners across 180 countries eligible to participate.
Organizers say most existing AI evaluation frameworks are built around a narrow set of dominant global languages, leaving Africa’s linguistic diversity — including widespread multilingualism, dialect mixing and culturally specific communication patterns — largely untested. They describe this complexity as a uniquely rigorous stress test for modern AI systems.
“The future of AI will not be defined solely in Silicon Valley or Beijing. It will be defined wherever AI meets linguistic and cultural complexity at scale,” said Celina Lee, CEO and co-founder of Zindi. “Through this challenge, we are positioning African AI talent at the centre of shaping global standards for trustworthy AI.”
Louis Powell, director of AI initiatives at the GSMA, said the collaboration addresses a critical gap as AI adoption accelerates across Africa’s mobile ecosystem. “Strengthening AI trust and safety is essential to unlocking the full potential of AI for inclusive digital growth,” he said.
