But have you ever hit the Closed Caption (CC) button on a YouTube video player? “Caption Actions” come up with ‘Transcribe Audio’ on top, i.e., convert the video’s audio into text.
Oh, they label it BETA (Like all of Google.com until ~2005), and they warn you specifically… “Please Note: Transcribe is an experimental service that uses Google’s speech recognition to provide automated captions for video.”
Well, when you use that audio transcription function, prepare for some laughs, or if your purpose is serious or your mission life-saving… red faces!
Here is a couple of especially egregious examples in the images below:
Ethan Zuckerman, in a wonderful TED Talk called “Listening to Global Voices,” lays out the current situation with the attempts to automate cross-cultural communications:
“I’m guessing that most of you don’t even speak Chinese—which is sort of sad if you think about it, as it’s now the most represented language on the internet. Fortunately people are trying to figure out how to fix this. If you’re using Google Chrome and you go to a Chinese language site, you notice this really cute box at the top, which automatically detects that the page is in Chinese and very quickly at a mouse click will give you a translation of the page. Unfortunately, it’s a machine translation of the page. And while Google is very, very good with some languages, it’s actually pretty dreadful with Chinese. And the results can be pretty funny. What you really want—what I really want—is eventually the ability to push a button and have this queued so a human being can translate this.”
Now, if you’re bi-lingual, and you really want to laugh, hit the “Transcribe Captions” button then “Translate Captions” for twice removed native language gaffes, malapropisms, and juicy double entendres… Who knows what sort of political trouble this could stir up?!?
And, to witness this occasional humor really being milked for laughs, here’s Rhett & Link’s “ULTIMATE Caption FAIL“…