User Collaboration at Scale [And the end of Extinct Languages?]

I was listening to 702 radio the other day and one of the callers was an African lady, who made an interesting point:

She said that she had looked everywhere in SA to find classic fairy tale stories (think Cinderella, Jack and the Beanstalk, et al) available in local languages like Xhosa and Zulu.

So this got me thinking two things: 1) how tragic it is for a parent, to not be able to tell these kinds of stories to their children in their mother tongue and 2) that there must be a way that technology can solve this problem.

Some Context

There are an estimated 1500 – 2000 different African languages! To put that in perspective, Nigeria alone has over 500 different languages! The sad part though, is that many of these African languages are slowly dying.

The concept of ‘extinct languages’ is a bizarre one to me because language (especially ones mother tongue) is often taken for granted. I could never imagine the english language being extinct… but for some African’s the sense is that their tribal languages are under threat… namely because if a language isn’t spoken at home, then it likely that is won’t be passed on to future generations.

And I can’t think of a better way to convey the power of a language… than through story telling.

I think the opportunity to translate not just your classic fairytale stories, but all books into a multitude of languages (especially African languages) is an inspiringly big idea.

The challenge then, is how come up with a solution that enables the translation of books like classic fairytales written in English to local African languages like Zulu and Zhosa and at scale… and for FREE.

Constraints

Before we tackle the worlds books, let’s start with just the scope of fairy tale books.

Naturally one would assume that it’s best just to get a bilingual to translate the books into vernacular but when one considers the number of fairytale books in play and the plethora of languages across the African continent to cater for, it is very unlikely that there would be enough bilinguals to do the translations.

So what if we could automate the translation using computers? Interestingly, we are about 10 -15 years away before we have artificial intelligence (AI) sophisticated enough in order to do accurate book translations and at scale.

Predictive text fail

So we need to think differently…

Insights

People really want to learn a new language – there are currently over a billion people who are currently learning the English language – and that is just English! So what if we could use this motivator to translate the books for us and maybe even more than just fairytale books… but all books globally.

The global book market is huge – it is estimated to be worth in excess of $100 billion. While sizeable, there are opportunities to expand the market by employing a market development strategy realised by selling existing books into adjacent markets (local language book buyers).

To learn a new language is actually expensive – this is especially true in developing economies. If we could simultaneously come up with a model that enables a free and fair model for language education that would be stellar.

The Business Model 

I’ve taken a stab at illustrating a proposed business model below:

Wikitranslate

Our solution would create a value exchange between book publishers, language learners and the wider public.

Book publishers – they would get the opportunity to increase book sales by selling translated books to the public.

Language learners – they would get the opportunity to learn languages by simultaneously translating books (note: would require a team computer scientists to work out exactly how this would work)

The public – they would benefit from getting the opportunity to purchase translated books.

Making it Happen

I think that this concept is quite interesting because it lends itself to massive user collaboration at scale.

In order for the model to work we would need to get all three audiences using the platform and perhaps the biggest challenge would be on how to work out the appropriate algorithms that would realise an accurate translation in a specific language.

Further consideration would also have to be given to the rights of book publishers i.e. to what extent do they have the rights to translate books which they publish?

In terms of access to bilinguals, this was of course a working assumption, but I think with the right go-to-market strategy and appropriate incentives… anything is possible. They may just be out there… and after all the power of the Internet can do amazing things.

So who knows? Maybe we could even prevent the extinction of a language or two…

 

,

No comments yet.

Leave a Reply