For a long time, the conversation around African languages has sounded like a eulogy. We’ve heard the warnings: indigenous languages are disappearing, young people are losing fluency, English and French dominate education and work, and eventually some of the languages spoken by our parents and grandparents could become languages we only remember hearing at home. But something interesting is happening.

A new generation of Africans is taking those same languages and putting them somewhere they have historically had very little presence: online. They are making TikToks in Yoruba. Teaching children Zulu through comedy. Creating digital media in Igbo. Building datasets in Hausa. Developing artificial intelligence that can understand African languages. Turning proverbs into social media posts and language lessons into entertainment.  And perhaps most importantly, they are making indigenous languages feel less like something we are supposed to preserve and more like something we can actually use.

That distinction matters. Because languages don’t survive simply because we have dictionaries for them. They survive because people use them to flirt, gossip, argue, joke, teach, make money and tell stories. Increasingly, young Africans are finding ways to do all of that online.

The internet isn’t necessarily killing African languages

Image Credit: Turkana Wildlife

The threat is very real: Africa is one of the world’s most linguistically diverse continents, but many African languages remain severely underrepresented in education, publishing, media and technology. The problem becomes particularly obvious when you look at artificial intelligence.

In a 2026 article for UNESCO, researcher James Maisiri described African languages as an AI blind spot. AI systems trained predominantly on dominant languages such as English can perform dramatically worse when asked to understand African languages. UNESCO cites research showing that ChatGPT recognises only around 20% of written Hausa sentences, despite Hausa being spoken by more than 80 million people in Nigeria alone. It is a strange contradiction. You can have tens of millions of people speaking a language and still have technology that barely understands it. But this is also where a new generation of African researchers, creators and entrepreneurs is pushing back. Instead of waiting for Silicon Valley to decide which African languages deserve a place in the future, they are beginning to build the infrastructure themselves.

And social media is where much of this cultural shift is becoming visible.

TikTok has become an unlikely language classroom

@lungileandluyanda

WATCH TILL THE END! Lenzan manje leli ninja 🙆🏽‍♀️🙆🏽‍♀️🤦🏽‍♀️ #mommydaughterduo #learningzulu #zulutiktok #tiktoksouthafrica🇿🇦

♬ original sound – LungileAndLuyanda

Language learning traditionally has a very specific image: textbooks, classrooms, conjugation exercises and someone telling you to repeat a phrase until you get the pronunciation right.

TikTok has decided to make it fun. In South Africa, Lungile Zenda became a viral example of this when she began posting videos of herself trying to teach her daughter, Luyanda, Zulu. The premise was simple: Luyanda was growing up in an environment where English dominated, and Zenda wanted to make sure her daughter could speak the language of her family.The videos were funny because learning a language is often funny. There were mistakes, corrections and the occasional struggle with pronunciation.

But people weren’t just laughing. They were relating.

Reuters reported that Zenda’s TikTok account grew to more than 150,000 followers, with her videos resonating with other South African parents worried that their children were losing proficiency in their mother tongues. The interesting thing here is the format. Zenda wasn’t presenting Zulu as a museum artefact. She wasn’t standing in front of a whiteboard explaining why linguistic preservation matters. She was simply trying to teach her child. And the internet turned that family exercise into a public conversation about language, identity and the generational gap. That is perhaps one of the most effective ways to preserve a language: make it normal.

Young Africans are making indigenous languages entertaining again

The same thing is happening in Nigeria. A 2026 study examining indigenous-language content on TikTok found that Yoruba made up more than half of the indigenous-language content analysed. Comedy was the most popular format, while subtitles were widely used, allowing people who did not completely understand the spoken language to follow along.

That combination is important. Because if you want young people to engage with a language, you don’t necessarily have to tell them to learn it.

Give them something they actually want to watch, and the language becomes the vehicle rather than the lesson.

Research into digital media and Yoruba has similarly found that platforms such as Facebook and Instagram are being used to circulate Yoruba proverbs, music, cultural knowledge and contemporary conversations. And this is where the internet becomes particularly powerful. Traditional media has always had gatekeepers. Getting a language newspaper, television programme or radio show funded requires money, infrastructure and an audience advertisers believe is commercially valuable. TikTok doesn’t ask for any of that. A teenager with a phone can decide that Yoruba, Igbo, Twi or Zulu is the language of their next video. And if people watch it, the language has suddenly occupied a digital space it might never have been given by traditional media.

Igbo is finding new life in digital media

The story of Igbo is particularly interesting. A peer-reviewed study published in Frontiers in Communication in January 2026 examined the impact of BBC News Igbo on the language’s status and use. The researchers found that digital media had helped increase Igbo’s visibility, revive interest, strengthen language pride and expand the situations in which people felt comfortable using it. 

The study is particularly useful because it moves the conversation beyond vibes. The researchers surveyed BBC News Igbo users and found evidence of what they describe as a domain expansion: people reported using Igbo in places such as social media, work, school and socio-cultural settings.

That might sound small. It isn’t.

A language becomes vulnerable when it gets pushed into fewer and fewer parts of life. If you only use Igbo with your grandmother but English everywhere else, English becomes the language of modern life while Igbo becomes the language of memory. Digital media can interrupt that process. Suddenly, Igbo can be the language you use to discuss an election, watch the news, debate football or scroll through your social feeds. It becomes a language that talks about the present. And that is one of the most important things a language needs to survive.

The next battleground is AI

Social media might be making African languages visible. AI could determine whether they remain technologically relevant. The problem is data. Artificial intelligence systems learn from enormous quantities of language data, but African languages have historically had much less digital material available than English, French, Spanish or Mandarin. A 2025 research project called NaijaVoices is trying to change that. Researchers collected more than 1,800 hours of speech from more than 5,000 speakers, focusing on Yoruba, Hausa and Igbo and deliberately including different accents, dialects, ages and speaking styles. This is much bigger than creating a translation app. Speech data is infrastructure. It can help build better speech recognition, transcription, voice assistants and other language technologies.

The researchers argue that African languages are significantly underrepresented in existing speech technologies and that this lack of data limits the ability to build reliable tools for African users. There are already smaller examples of this work happening elsewhere. In 2026, researchers published work on text-to-speech technology for Efik, creating an initial speech corpus and testing several AI models. The project demonstrated that even relatively low-resource African languages can begin to enter the world of speech technology, although the researchers also found persistent challenges around tone and pronunciation. (arXiv⁠) This matters because technology changes what we consider a “useful” language. If your language can be spoken to a phone, transcribed by a computer, understood by an AI assistant and used to access information, it has a place in the technological future. If it can’t, speakers may increasingly have to switch to another language to interact with the technology around them.

Preservation doesn’t mean freezing a language in time

Image Credit: Afrique Noire

There is, however, a danger in the way we talk about preservation. Sometimes it sounds like the objective is to put African languages in glass cases and keep them exactly as they were. That isn’t how languages work. Languages change. They absorb slang. They borrow words. They develop new expressions. They respond to music, politics, migration and technology. The Yoruba spoken by a teenager in Lagos today does not have to sound exactly like the Yoruba spoken by their grandparents for it to remain Yoruba. In fact, trying to force young people to speak an artificially “pure” version of a language may make the language less attractive to them.

The more interesting question is whether African languages can continue to adapt.

Can Hausa be used to explain cryptocurrency?

Can Zulu exist naturally in memes?

Can Efik work with voice assistants?

Can a young African in London speak their grandmother’s language on TikTok and feel like they are participating in contemporary culture rather than performing heritage?

Those are the questions that matter now.

The work is not without problems

Putting a language online doesn’t automatically save it. AI can mistranslate. Social media can encourage code-switching rather than fluency. Algorithms may favour English-language content because it has a larger audience. And digital platforms can still struggle to correctly identify African languages. A 2025 study examining language detection in digital spaces found that African languages such as Yoruba, Kinyarwanda and Amharic were often poorly represented in conversational online data, with code-switching making it harder for language-detection systems to accurately identify them. The researchers found that professionally produced local news content provided much cleaner language data than conversational social media. 

There is also the issue of dialect. One language can contain enormous regional variation, and building a technology that recognises one dominant version doesn’t necessarily mean it works for everyone. That is why the work being done by African researchers is so important. The goal shouldn’t simply be to make African languages visible to technology. It should be to make technology capable of understanding Africans as they actually speak.

The real revolution is making the language useful

Perhaps this is what separates the current generation from previous language-preservation efforts. They aren’t necessarily approaching the issue with the language of preservation. They are approaching it with the language of culture. A creator makes a funny video in Yoruba because that’s what feels natural. A mother teaches her daughter Zulu because she wants her to communicate with her family. Researchers collect thousands of hours of Nigerian speech because they don’t want AI to leave Nigerian languages behind. A news organisation publishes in Igbo because people want to consume information in a language they understand. None of these things necessarily begin with the grand ambition of “saving a language.” But collectively, they do something more important.

They keep the language in use. And perhaps that is where the future of African languages will be decided.