Strong Language

New voice transcription feature in Google Docs censors (some!) swearwords

Google Docs announced today that you can now create documents using your voice.  And of course, like any good linguist, I immediately went to try to stump it. It’s pretty good, actually — it recognized both pronunciations of “gif” and “aunt” in the contexts “animated ___” and “uncle and ___” although it tended to assume that I might have the bit/bet merger, which I most emphatically do not, and thus presented me with a few transcriptions that felt like odd candidates to me.

But then I tried swearwords and hit the fucking jackpot. Here’s a screenshot of what I got:

Fuck, shit, and cunt are bleeped out with asterisks — well, if you’re going to censor swears, that’s the least surprising. God damn, hell, bastard, and bitch aren’t bleeped — well, they’re on the tamer side. Tits and piss aren’t bleeped either, but it took several tries to convince the voice transcriber that I actually wanted to say them — those “f”s you see are me trying to say piss, which it only finally accepted once I added off (fuck is not helped by the addition of off, and motherfucker, interestingly, gets its first part transcribed just fine but as a separate word). Similarly, I eventually got it to write tits, but it aimed for text and tents first. And perhaps most interestingly, cock and ass are uncensored, but cocksucker and asshole get bleeped. Surely a sociology study could be written about such a distinction.

I mean, I get the point, surely you don’t want your voice transcription program accidentally peppering your business memo or school essay with f-bombs, but it does seem like certain fuckers might want the option — it would certainly make it hard to draft a post for this blog! I poked around in settings for a bit but I couldn’t find a “disable SafeTranscribe mode” akin to “disable SafeSearch”, and that does seem like a fucking loss!

Anyway, this was just a quick list of the swears I could think of, aided by the 7 words you can’t say on TV, so if anyone thinks of other words to try, you can enable this feature in the Tools menu of any Google Doc, and do let us know in the comments whether or not they’re bleeped! Also, anyone know what the hell earlier dictation programs did with swears?