accents, identity and techno-utopianism in the era of dumb ML

Source: David Knickerbocker

0. Context to context

Throughout last year, I worked with the National Technical Institute for the Deaf on a new metric, ACE2, that evaluates the efficacy of automatic speech recognition software. Speech recognition products have traditionally been evaluated using a metric self-explanatorily called the word-error rate (WER), but it has been found to be have…