Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
Youtube

YouTube's Captions Insert Explicit Language In Kids' Videos (wired.com) 78

An anonymous reader quotes a report from Wired: A new study (PDF) of YouTube's algorithmic captions on videos aimed at kids documents how the text sometimes veers into very adult language. In a sample of more than 7,000 videos from 24 top-ranked kids' channels, 40 percent displayed words in their captions found on a list of 1,300 "taboo" terms, drawn in part from a study on cursing. In about 1 percent of videos, the captions included words from a list of 16 "highly inappropriate" terms, with YouTube's algorithms most likely to add the words "bitch," "bastard," or "penis." Some videos posted on Ryan's World, a top kids' channel with more than 30 million subscribers, illustrate the problem. In one, the phrase "You should also buy corn" is rendered in captions as "you should also buy porn." In other videos, a "beach towel" is transcribed as a "bitch towel," "buster" becomes "bastard," a "crab" becomes a "crap," and a craft video on making a monster-themed dollhouse features a "bed for penis."

Automated captions are not available on YouTube Kids, the version of the service aimed at children. But many families use the standard version of YouTube, where they can be seen. Pew Research Center reported in 2020 that 80 percent of parents to children 11 or younger said their child watched YouTube content; more than 50 percent of children did so daily. [...] YouTube spokesperson Jessica Gibby says children under 13 are recommended to use YouTube Kids, where automated captions cannot be seen. On the standard version of YouTube, she says the feature improves accessibility. "We are continually working to improve automatic captions and reduce errors," she says.
"The team also ran audio from kids' YouTube videos through an automated transcription service offered by Amazon," adds Wired. "It too sometimes made mistakes that made the content edgier. [...] 'Fluffy' became the F-word in the transcript of a video about a toy; one video host asked viewers to send in not 'craft ideas' but 'crap ideas.'"
This discussion has been archived. No new comments can be posted.

YouTube's Captions Insert Explicit Language In Kids' Videos

Comments Filter:
  • YouTube's automatic captioning system gets it right most of the time. Every now and then, the AI tosses in a naughty word for shakes and giggles. If you're a content creator, you do your own captions or pay a third-party company to create captions for you.
    • by shanen ( 462549 )

      Legitimate complaint, though I don't understand the low Score. "No comment history available" means what?

      As for the story, I'm hard pressed to care less. However I have a number of beefs with YouTube. The new one involves the ads. I think they've started a new scam. When the video ads are supposed to play, it just sits there doing nothing. Then, after 10 or sometimes up to 30 seconds, the video starts playing. It's almost as though YouTube is trying to avoid my cancel-ad clicks while telling the advertisers

      • "No comment history available" means the post has not been moderated yet. No positive or negative mods yet and no karma bonus points.

        I have also seen YouTube videos that appear to just sit there for 10 to 30 seconds but I attribute that to my ad blocker hiding the ads from me. YouTube waits until it thinks the ad has finished then proceeds to display the video. Since I don't see the ad, I can't click the "Skip ad" button so I have to wait.

        If you have an ad blocker, you can test this by temporarily turnin

        • by shanen ( 462549 )

          No ad blocker. I basically accept the deal and I know there's no free lunch.

          On your last suggestion, I have a similar strategy by switching to a different video.

          I would like to estimate the advertising overhead more accurately, but pretty sure it's lower than on television in the days when I used to watch commercial TV. I pretty much quit watching TV before cable, so I can't assess the advertising there, but what boggles my mind is noticing the advertising channels (in passing). Who in their right mind woul

          • by I75BJC ( 4590021 )
            During the pandemic, I watch a lot of YouTube videos. (I prefer to control what I see and dislike commercials.) After months of commercials and the blank spots, I subscribed to YouTube Premium (?). No commercials and more enjoyment. The blank spots are gone to0. The only "commercials" are the sponsored-by announcements by the video host. I can listen or fast-forward. I always use automatic English captions. YouTube gets this right most of the time but there are mistakes at times. Generally, this ha
            • by shanen ( 462549 )

              No way for me (though I agree with your filtering considerations). I see that option as paying the blackmailer. YouTube is fundamentally a criminal enterprise, the google is increasingly EVIL, and the main thing I want is alternatives. But the game is rigged to prevent sincere competition from competitors.

              But you are still being subjected to the in-video advertising and pitches for their sponsors, so even that part of the deal is bogus.

    • YouTube's automatic captioning system gets it right most of the time. Every now and then, the AI tosses in a naughty word for shakes and giggles. If you're a content creator, you do your own captions or pay a third-party company to create captions for you.

      In my opinion, YouTube should require closed captioning on every new video that gets posted. The content creator can outsource the job to a third party or use YouTube's voice recognition algorithm to create a first draft, then manually correct the captioning. Requiring closed captions on new content won't fix the millions of videos already in YouTube, but it should be mandatory for all new content.

      I have a deaf friend that tries to use YouTube for everything from home repairs (how to replace a toilet flap

  • holy shit (Score:3, Funny)

    by Anonymouse Cowtard ( 6211666 ) on Friday February 25, 2022 @10:46PM (#62305221) Homepage
    This is seriously fucked up. Little fuckers should live in glass fucking domes. I feel sorry for the fucking parents. The poor cunts have to be fucking helicopters to monitor their little fuckers' every exposure to the real fucking world.
    • by fermion ( 181285 )
      What kind of messed up world is penis a taboo word. What is next. Kidney. Tibia? God forbid that our kids be educated and have a full vocabulary.
      • That'd be a Religious world.

        ("Conservative" in the USA)

      • There are situations in which using the word "penis" when talking to a child would be appropriate. Vanishingly few of those situations seem reasonable to handle in a Youtube video. If your Youtube content aimed at children is talking about penises, you're probably doing something wrong.

        • by fermion ( 181285 )
          I am so glad this was not the metric used when I was growing up. I would as stupid and uncultured as everyone else. In church, which fortunately was largely populated by academics, these were the words, and you spoke like an educated person, not like a hick
  • Dick and Muffy have fun with her fury belly!
    Dogs love to have their belly rubbed.

  • Youtube captioning system correctly identifies subliminal messagery in videos targeting kids.
  • by davmoo ( 63521 ) on Friday February 25, 2022 @11:17PM (#62305245)

    That's what happens when parents let the kids use the adult version of a product. It is in no way Youtube's fault.

    • YouTube Kids is trash and is full of junk content. The only useful option is full YouTube with parental supervision.

    • That's what happens when parents let the kids use the adult version of a product. It is in no way Youtube's fault.

      On one hand, "I use a machine as a surrogate parent and it's a worse parent than I am" isn't sympathy-invoking.

      On the other hand, the sooner your kids are desensitized to profanity, the sooner you can not give a shit about what the subtitles are doing.

      On the gripping hand, this might encourage them to read, since they might find something juicy in the subtitles.

      • by Entrope ( 68843 )

        The summer of 2020, my kids spent a lot of time at the neighborhood park with a babysitter. There was graffiti there briefly, but someone painted over it fairly promptly.

        I thought very little of it until, a few days later, I was struggling with something in the kitchen and said "stupid jar" (or whatever). My almost-six-year-old, dutifully and very matter-of-factly, chimed in from the next room: "'Stupid' is a bad word. But not as bad as 'fuck'."

        He wasn't wrong....

        • The summer of 2020, my kids spent a lot of time at the neighborhood park with a babysitter. There was graffiti there briefly, but someone painted over it fairly promptly.

          I thought very little of it until, a few days later, I was struggling with something in the kitchen and said "stupid jar" (or whatever). My almost-six-year-old, dutifully and very matter-of-factly, chimed in from the next room: "'Stupid' is a bad word. But not as bad as 'fuck'."

          He wasn't wrong....

          I hear you. The real problem with kids and profanity is that other people get offended. You don't want them saying "shit" around people who will be bothered. I mean, it's not great to just sprinkle them constantly into discourse, but these words really do only have the power over us that we let them have.

        • by _merlin ( 160982 )

          My goddaughter and her dad both like to remind me that I taught her the word "fuck" (not literally taught her, she overheard me getting angry at someone while I was on speakerphone).

  • It's like machine translation of foreign texts. Sometimes, the results are humorous, some times they're offensive, and sometimes they get it right.

    They always say they hope to improve it some day.

    Maybe it should default to "off" more often?

    • by gweihir ( 88907 )

      I recently recorded a lecture I was giving in MS Teams. The "transcript" was the most appalling nonsense I have sever seen. Of course, the lecture was in German, but Teams apparently did not notice that and was trying really hard to transcribe all of it in English. Will show that to my students when I have a minute as example of how broken and unfinished some "modern" tech actually is. Of course, when it does not even recognize that a different language is spoken, it will make the most horrible mistakes eve

      • by burni2 ( 1643061 )

        I so do hope that Putin nor Biden use such shitty translators .. well otherwise we will all get a serious sun burn.

        • by sjames ( 1099 )

          Notably, Khrushchev DID say "We will bury you", but it wasn't meant as a threat, just his belief that the Soviet Union would still be there after the United States collapsed in on itself.

          Even correct literal translations can go very wrong.

      • Reminds me of a Google Voice transcript of a spam voicemail that was partly in Chinese. The last "sentence" was "I do not cranking I soon 2013 tonight."

        • by gweihir ( 88907 )

          Reminds me of a Google Voice transcript of a spam voicemail that was partly in Chinese. The last "sentence" was "I do not cranking I soon 2013 tonight."

          Yes, pretty much. The only thing this is good for is getting laughs.

      • "Teams defaulted to the assumption that the spoken language was English" is hardly justification for the broad claim that modern tech is "broken and unfinished". You suggest that Teams should "recognize that a different language is spoken" as if that's a trivial thing to do.

        Did you even try to change Teams to a "German" mode? If not, this seems analogous to me to complaining that your computer blew up because you plugged it in in Europe while the PSU had the input voltage switch et to 110 instead of 220

        • If you're getting a bunch of low confidence metrics on best matches, run it against other languages, and if you find one scoring far higher, well there you go. Not trivial, but not hard compared to building a speech to text system to begin with.
          Google has that option. [google.com] Detecting a single language is easier than what that's doing; that's detecting multiple languages in the same source.
          • Yeah, yeah. I fundamentally agree with OP that it'd be a nice feature for Teams to recognize the language you're speaking. And yeah, I know it's possible. But find the complaint whiney nonetheless.

            "Ugh! This app is interpreting the things I'm saying (through a cellphone mic, in a lecture hall, spoken at conversational speed) and it doesn't even detect that I'm not speaking the language that the app is set to! It should totally suggest changing to another language that it detects!! :'-(. Unusable! 1 out of

            • by gweihir ( 88907 )

              Obviously you have not even a basic clue how speech recognition works and how badly they must have messed up the parameters for this to even be possible.

        • by gweihir ( 88907 )

          "Teams defaulted to the assumption that the spoken language was English" is hardly justification for the broad claim that modern tech is "broken and unfinished".

          A speech "recognition" software that cannot tell it is listening to a different language than it expects is.

  • by OrangeTide ( 124937 ) on Friday February 25, 2022 @11:50PM (#62305291) Homepage Journal

    I wonder if the offensive language on YouTube is worse than what kids hear at home from their own parents?

    • by cobbaut ( 232092 )

      Exactly.

      Also, why is "penis" an offensive word? It is a body part, are kids not supposed to know this exists?

      • The common thinking now is they absolutely should know body parts. Because 'Daddy ate my cookie' can have very different meanings if you're teaching cutesy words to refer to their genitals.
      • by kmoser ( 1469707 )
        Also, why is bitch offensive? It means female dog. "Bastard" means someone born of unmarried parents. Nothing offensive about that.
        • The second dictionary definition of bitch is "a malicious or promiscuous woman," although my understanding is that it means "an incessantly complaining woman." It's insulting. If you're picky about such things, it's also sexist.

          Shakespeare may be responsible for the bad connotation of the word "bastard."

  • Or not. Because they all will have heard and read far worse. Seriously, stop treating kids as precious little flowers that die on the first whiff of something "inappropriate". Any kid with a smartphone will have seen hardcore porn, people getting killed and really disturbing things like Trump speaking. For most this will cause absolutely no problem at all and the rest cannot really be isolated from this either.

    • Even grown up people who get paid to see people getting killed get PTDS (look up how facebook moderators feel about their work). Those sorry sods are doing it not because Zuk is a sadistic bastard that wants pay people to suffer. They do it because other FB users don't want to be traumatized as well. Apparently you look think that what grown ups can't handle is totally ok for children.

      About swearwords, it's not like hearing them in isolation will traumatize children. It's because parents don't want to hear

      • don't want to hear their child say "Happy 50th birthday you fucking cunt!"

        How do you teach them not to say "Happy birthday, poopy-head!" The aforementioned phrase is mean. When the kids see it, you explain it to them that way. Another lesson learned and your kids will be better for it.

      • by gweihir ( 88907 )

        Apparently you have your head deeply lodged in your backside. Screening stuff for work is fundamentally differently from looking at stuff you selected yourself and can stoop looking at any time.

  • by Alsee ( 515537 ) on Saturday February 26, 2022 @02:53AM (#62305531) Homepage

    Our Earth is degenerate in these later days; there are signs that the world is speedily coming to an end; algorithms are feeding our children pornography; bribery and corruption are common; children no longer obey their parents; every man wants to write a book and the end of the world is evidently approaching. <ref>Assyrian clay tablet, 2800 B.C</ref>

    -

    • by burni2 ( 1643061 )

      .. algorithms are feeding our children pornography; bribery and corruption are common; ..

      You really should stop smoking shitty shit shit it makes you delusional, and you should start a healthy off-internet diet because your mind is poisoned.

      Some shitty reference to bible or apocryph or some other text written by an ancient shit smoker who was keen on seeing the world ending.

      However the end could really be near, I mean when Putin crosses the NATO border he is ultimately outgunned and outmanned (population vs. weapons vs. the painful effect of western weapons russian soldiers are experiencing now) and even with only the european forces

      And well with 90% of Ukrainian people (even people of russian decent) lusting for revenge in his back (along with people from Georgia and many other occupied regions that will rise when the military is shifted towards NATO borders)

      And Putin's ego will drive him to press the red button .. and well perhaps younger people don't realize what a nuclear 3rd world war will really mean - we will soon all be dead.

      Now you can go crazy.

  • What kind of nanny state do we live in where there are 1300 words considered “taboo”, whatever that means. Is there a fleet of Sunday-school teachers making up this list. Seriously, I get to about word number 50 in my list of worst coarse language and I’ve fallen into the realm of “mildly offensive to the thin-skinned”. With an average vocabulary of 20-30 thousand words, this is 4-7% of our words. Sure glad I don’t have a host of risk-averse corporate advertisers my life

    • And just how many of the 13 languages that YouTube auto captions are available in are you fluent in? Whilst I know a few less than polite words in French, German and Spanish, quite honestly, I couldn't take a guess at how many words are taboo in Vietnamese, Korean, Japanese or Indonesian.
    • Did you count slurs? Everyone's familiar with the n*r but there's a long, long list of slurs for every ethnicity and national origin there is. I'm of Italian descent; was dago and wop on your list?
  • Is that kid's content creators are putting out explicit content to children designed to bypass the mental radars of parents. If it can fool YouTube's automatic-CC algorithms, it can absolutely trick a developing mind into hearing a swear, and was likely intended to do exactly that.
  • You think this will prevent kids learning bad words? Have you seen an adult who grew up not knowing them?

  • ... "bed for penis."

    It's not just the FCC [youtu.be] attempting to censor anatomy.

  • Since when is penis an explicit word rather than the anatomically correct description of a body part. I can think of many variations of the word penis which would be explicit but I think this says more about the writers of the article than it says about Youtube.

  • The YouTube captioning algorithm also tries to caption Dutch and German ads/commercials resulting in a really unreadable mess.
    My main pet peeve with YouTube commercials/ads is that the volume is way louder than the video and so my mute button on the TV remote needs to pressed directly after selecting a video or otherwise the neigbours will complain....

  • Thank god that all this is backed up by the mountains of scientific evidence that support that children reading the word “penis” has negative effects on them, and goes into great detail describing what those negative effects might be.

  • Subject line says it all.

    I can't say that I've noticed YouTube having a "captioning" system, but maybe I just use it differently to the average user. Concentrating on the lectures I download, that sort of thing. Do people use it for casual entertainment too?

  • "words in their captions found on a list of 1,300 "taboo" terms, "

    Most teachers would love if their pupils had a 1300 word vocabulary.

  • The word "penis" is highly inappropriate so what should we say? Cock?
  • People who do nothing find ways to get offended by finding minor glitches in what is otherwise an exceptionally good result obtained by people who do something.
  • "Automated captions are not available on YouTube Kids, the version of the service aimed at children. But many families use the standard version of YouTube"

    Or maybe two sentences:

    "one video host asked viewers to send in not 'craft ideas' but 'crap ideas'."

    And I guess this "study" was one of them. Maybe also "study" some Quentin Tarantino films and conclude that children are being "exposed to a lot of violence" because, although those films aren't aimed at children, some films are, and some parents might have

  • There are 1300 "taboo" terms? Really? Let's see the list.

  • and you let your kid on youtube, you're a shit parent and deserve whatever you get

  • In olden times, while walking on the uphill both ways path from or to school, our flirtations with forbidden speech were mostly trying to say "my father works in a shipyard" while holding the tongue down with a popsicle stick. We made a living in the youth stratum by exchanging pennies, nickels, half-sticks of gum, and small and somewhat perishable frogs.
  • People should learn to proofread their postings before clicking on the send button. I can't say "do no evil Google" isn't responsible but a lot of adults do enjoy using swear words to emphasize a point which could allow the word-guessing fill-in-the-blanks software to decide what you meant to say.
  • The captions are not "inserting" anything "into" videos owned by kids. The captions system is an overlay, on top of the video, that generates captions automatically based on what it hears -- it isn't put "into" the video; it's just on top of it in a z-axis sense. The original video file is unaffected, and can be disabled, and isn't even on the kids section of Youtube. It happens across all videos of any demographic, and not just on YouTube; it happens on Twitch also, and conceivably anywhere auto-generated
  • Why is penis a bad word? Is a liver a bad word too? How kids should talk about their penises? Wee-wee?

The trouble with being punctual is that nobody's there to appreciate it. -- Franklin P. Jones

Working...