Automated speech recognition systems that convert spoken language to text have been shown to perform better listening to white voices over Black voices.

Was this research made exclusively based on data from English speakers from USA?

I’m asking that because English dialects (yup, plural) used there by black people are noticeably different from the ones used by the others, in a way that is outright alien even in the rest of the New World. And it would help to exactly which people are or aren’t being excluded by this technology, thus being important to mitigate the exclusion.

I know that I’m commenting on a small detail, but that caught my attention.