Projekte

Al is fed by existing language data, so it will do no more than hold a mirror up to what is already out there. But it’s worth pointing out that the linguistic character of that mirror will be influenced by the location of the large companies investing in Al. lf most of these are American, for example, we will see more American English (spelling, punctuation, vocabulary, grammar) in contexts where traditionally other dialects would be usual. And it remains to be seen whether other languages than English will develop an Al presence comparable to what we have seen so far in the internet, which has become an increasingly multilingual medium over the past 30 years. When the internet began, the only language with a significant presence was English. That language is still the dominant voice […] but around 500 languages have a noticeable presence now, and perhaps another thousand are dipping their linguistic toes into the internet water, including (especially in social media) several minority and endangered languages. But with over 6000 languages in the world, and an appreciable but unknown number of dialects, there is a long way to go before the electronic world replicates the linguistic diversity of the planet.

Professor David Crystal in JLLT Volume 14 (2023) Issue 2 (Journal of Linguistics and Language Teaching)