Goloka Analytics/Meta partnership: What this means for languages and AI development in Nigeria


 

By: Quadri Yahya

Goloka Analytics, a spin-off of Dataphyte, recently announced its partnership with Meta on a multilingual Artificial Intelligence project.

The Nigerian technology company said it will “collect data in five Nigerian local languages for the training of Meta's Multilingual Al models and to promote digital inclusivity worldwide.”

A press release announcing the partnership stated that the initiative is part of Meta's Local Language Partnership Program.

“Through this collaboration, Goloka Analytics, a technology company focused on point-of-interest data and insights using native intelligence, mobile, and geospatial tools, is contributing to an Al ecosystem where data is shared intentionally, owned by communities, and made to serve their interests”, Joshua Olufemi, founder of Goloka Analytics said.

Through this partnership, GOLOKA™ will contribute data in Yoruba, Hausa, Igbo, Fulfulde, and Urhobo.

AI and Nigerian languages

With AI's role as the technology of the future, tech companies are investing in multilingual models to enable localisation. For example, in 2024, Google added three Nigerian and 12 African languages to Voice Search and Translate Dictation, a development attributed to advances in AI.

Although Meta's new partnership with the Nigerian tech company follows Google in its focus on major Nigerian languages, it is a step further in including Nigerian languages in AI systems. Nigeria is home to more than 500 languages.

As the founder of Goloka Analytics explained, the core aim of language inclusion in AI models is “to strengthen the presence of indigenous Nigerian languages in AI-driven speech recognition and machine translation systems.”

"By applying our hyperlocal data collection model, we intend to provide high-quality linguistic datasets that help build more accurate and inclusive technologies. This effort aligns with our commitment to democratising data and supporting the use of indigenous languages in digital spaces. We are focused on collecting data that reflects local realities and promotes fair representation for communities that have long been underrepresented in digital technologies," he said.

What this means is that speakers of local languages like Urhobo, for example, will be able to use and access AI systems in a better way. "Our goal is not just to build AI systems about people, but to build them with people," Mr. Olufemi said.

The partnership will also engage academics, language speakers, and data collectors. “As this project unfolds, Goloka and Dataphyte invite stakeholders from the public, private, and civic sectors to join in shaping a future for Al that is inclusive, decentralised, and grounded in local realities”, the statement reads.

 

Languages and digital technologies

Linguists have noted the critical role of language in closing the digital gap in Nigeria and Africa. According to them, accessibility and inclusivity should mean that tech companies include local languages in digital platforms, including social media and AI systems.

Although some Nigerian and African languages have been included in digital technologies, many endangered, undocumented languages have yet to find a place.

A Nigerian writer, translator, and AI language data annotator, Abdulrosheed Fadipe, notes that some languages, such as Yoruba, Igbo, and Hausa, have been substantially included in digital technologies, while hundreds of others are yet to be documented or included.

“If languages are not included in the development of these tools, hundreds of millions of Africans may not benefit from the limitless opportunities offered by AI”, Abdulrosheed Fadipe had written in an earlier article.

Fadipe acknowledges that “the problem with minority languages is that there is not enough data to train AI models”, while noting that if “Goloka™ (and tech giants) want to do more, they should focus on minority languages.”

Language researchers have a huge role to play in this.

 

Nigeria efforts

While Nigeria has been projected to lead in the AI revolution in Africa, language is at the core of this. Nigeria has employed the services of a Nigerian tech startup, Awarri, to build the first Nigerian government-supported Large Language Model.The Nigerian government aims to use LLM, which powers AI models, to improve the representation of low-resource languages in AI datasets. 

This new development is part of the Federal Government’s push for digital transformation, focusing on the Internet of Things, AI, and blockchain technologies.

It is expected that this will boost employment opportunities in the country, which has millions of unemployed youth. The recent partnership between Goloka Analytics and Meta, as well as efforts by other tech companies and the government, will boost Nigeria's economy, linguistic and cultural diversity.

Post a Comment

Previous Post Next Post

Contact Form