With the accelerated advancement in generative artificial intelligence, the need for inclusive AI has never been greater.
This was one of the takeaways during a panel discussion at the recent demo day of Lelapa AI’s Vulavula natural language model solution.
Hosted under the theme “The future is today – your adventure of discovery”, the demo day brought together members of the local AI community, investors, innovators, industry leaders and young developers to explore Lelapa AI’s AI language technology.
Despite AI language technology making significant strides, it was agreed this technology revolution still marginalises African languages.
Among the key challenges noted is collecting data in African languages, given the costs involved and the quality of existing resources.
As a result, the panellists agreed there must be concerted efforts to shape these technologies, as well as build sustainable solutions and collaborate with local communities to build effective AI models, rather than remain passive consumers.
“As AI continues to advance, we must ensure African languages are not left behind,” said Pelonomi Moiloa, co-founder and CEO of Lelapa AI.
“By 2050, Africa will have the highest population, largely young and digitally connected. We need to build these technologies ourselves to serve our communities. It’s time for Africa to move from being a supplier of critical AI resources, like lithium and cobalt, to being a creator of cutting-edge technology solutions.”
Started in 2022, Lelapa AI is a research and product lab founded out of the need to address how AI can be used for solutions and applications from an African lens. It develops speech recognition tools for African languages.
The start-up’s founding members include Moiloa, Jade Abbott, Vukosi Marivate, Benjamin Rosman, Pravesh Ranchod and George Konidaris, all with backgrounds in academics, research, data science and engineering.
The release of natural language processing tools has amplified calls to dismantle some of the barriers when it comes to the development of these AI tools for African audiences.
This is spurred by the ethical challenges that have arisen as a result of AI’s rapid advancement. Notable among these is bias in algorithms, privacy infringements and more recently language, where it’s feared AI’s rapid advancement will have the potential to leave much of Africa’s 1.2 billion population behind.
The African continent is a multilingual melting pot, with as many as 3 000 languages spoken.
For example, South Africa has 12 official languages, but only one in 10 South Africans speak English at home – the language that dominates the internet.
To kick-off, Lelapa AI went to market with its first natural language model solution, named Vulavula, offering transcriptions, text analysis and translation capabilities.
Vulavula, which means to talk in xiTsonga, is an application programming interface that provides the capability to do multilingual communications for a business’s different applications, Moiloa told ITWeb after the event.
Its transcription tool is available in Afrikaans, isiZulu and Sesotho, according to Moiloa, describing it as the tool that sets the company apart from competitors.
“Our translation tool is across multiple languages, including South African, West and East African languages, as well as other language functionality for things like conversations – to a limited extent being able to identify what the intent behind somebody’s message is.
“Speech to text we also have, which is predominantly for local languages. We’re spanning eight languages, which is across East, West and Southern Africa. However, these aren’t all at the same level in terms of all of the features we offer at the moment.”
During the demo day, attendees were given the opportunity to interact with and test Lelapa AI’s technologies, including the newly-released InkubaLM natural language processing model.
Four primary activation zones were set up: the transcribe and leaderboard zone, learn a phrase zone, the InkubaLM zone and start-up zone.
The learn a phrase zone allowed attendees to engage with language in a fun and educational way, using Lelapa AI’s transcription technology to master phrases in various African languages. Meanwhile, the InkubaLM zone illustrated its role in advancing language preservation and digital inclusion.
The start-up zone showcased young developers using Lelapa AI’s Vulavula technology to build innovative applications. The space featured a combination of demo videos, static displays and roaming interactive demos accessible via mobile, highlighting the versatility and real-world potential of the technology.
Start-ups showcased included Clinic Bot, Voice-to-Accounting, Lethabo Galactic Bot, Bantu Keyboard and Insurance Quote Bot.
According to Moiloa, there are a lot of projects in the pipeline, with the main one being getting Lelapa AI’s technology into the hands of people through digital products and services offered by other businesses.
“This is important, not only because we need to prove the language is capable. We don’t need to look elsewhere in order to utilise the technology, but to help change the narrative around us to be able to do stuff that we’ve never been able to do before.
“We’ve built the technology and it’s there, but the next step is to have it taken up en masse, so that we can see this future that we’ve been promising ourselves for a while.”
Share