Is autocorrect racist? The campaign to get tech giants to stamp out ethnic name bias on phones, computers

As we’ve all encountered, frustrating autocorrect issues have been around since it was first introduced decades ago, with people venting their frustration for just as long (check out #damnyouautocorrect). 

This article is an interesting angle on it: Is autocorrect racist? The campaign to get tech giants to stamp out ethnic name bias on phones, computers

Adding a spreadsheet of names only partially patches part of the issue, because the problem is not just restricted to names.  If I were Microsoft, I’d be responding that spell checking and autocorrect is hard using traditional programming code, but we now have a new technology that can come to the rescue. I.e. Large Language Model Artificial Intelligence is going to literally re-write how autocorrect is implemented. LLM AI (in Microsoft’s case, almost certainly in the form of Microsoft's “Co-Pilot” AI) will revolutionise autocorrect. Rather than just matching against known correct spellings, LLM AI will be able to consider what someone types in the context of understanding the whole sentence. Autocorrect will never be 100% perfect though, given that, by its nature, needs to work before the sentence is completed.

THIS POST WAS CREATED IN FULL BY A HUMAN, WITHOUT USE OF AI.

#ORGANICALLYWRITTEN