The Federal Government of Nigeria has unveiled N-ATLAS, an open-source multilingual and multimodal large language model designed to support Yoruba, Hausa, Igbo, and Nigerian-accented English.
Dr. Bosun Tijani, minister of communications, innovation, and digital economy, announced the launch on his official X account Saturday during the 80th United Nations General Assembly.
N-ATLAS was developed to prioritize African voices in the global AI landscape, with a focus on preserving and advancing local languages. “By starting with Yoruba, Hausa, Igbo, and Nigerian-accented English, N-ATLAS emphasizes the importance of African voices and diversity in the foundation of AI,” Tijani said.
Equipped with advanced speech-technology capabilities, N-ATLAS includes language-specific automatic speech recognition models. These enable transcription, improve accessibility, and support local-language applications, from chatbots that provide government services to tools that power call centers.
The model can also transcribe radio, television, and online videos into text, generating captions and subtitles in the supported languages. It can summarize interviews conducted in local languages, detect user intent, and create automated responses tailored to Nigerian-accented speech.
Developed by the National Centre for Artificial Intelligence and Robotics in collaboration with Awarri Technologies, N-ATLAS marks a significant step in Nigeria’s broader strategy to position Africa as a key contributor to the future of artificial intelligence. Full documentation of the model is available on Hugging Face.
