Closed captioning (CC) based on voice recognition software (VRS) has certainly improved a great deal since it was first introduced -- I remember seeing lots of errors in days gone by -- but it ain't quite perfect yet, folks.
Case in point:
Earlier today I saw a report on an all-news channel about today's grand jury indictment of the current mayor of New York City, Eric Adams. As part of the report, the station showed an older video clip of the mayor saying "Make no mistake, when I was elected I promised to lead this city, and I will lead this city." Across tbe bottom of the screen the closed caption said "Make no mistake, when I was elected I promised to leave this city, and I will leave this city."
This is the stuff of which conspiracy theories are made, in my opinion. The people who depend on closed captioning to get their information received, in this case, a message completely opposite from the one the rest of the population heard.
Either voice recognition software needs more work or artificial intelligence (AI) isn't what it is cracked up to be, or perhaps both statements are true.
Am I right or am I wrong? Feel free to agree or disagree in a comment.
Closed captioning errors irritate me, too. Perhaps people need to articulate more precisely.
ReplyDeleteIn the early days of speech recognition, we had a lot of laughs at errors like that. One rather artificial example someone invented was "This nudist play can wreck a nice beach" (this new display can recognize speech). However, now I have to rely on voice readers, I have to say that the reader in MS Edge is very good. That's not captioning I realise.
ReplyDeleteI often use closed captions to avoid bothering people with noise and I notice there are many errors. I can usually pick the error by context but you make a great point. It surely skews the message and sometimes there's no way to correct that
ReplyDeleteSorry but this gave me a chuckle.
ReplyDelete