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

 



Forgot your password?
typodupeerror
×
Microsoft AI

Microsoft's AI Journalists Confuse Mixed-Race Little Mix Singers on MSN Homepage (theverge.com) 76

Microsoft's decision to replace human journalists with AI to run its news and search site MSN.com has been criticized after the automated system confused two mixed-race members of British pop group Little Mix. From a report: As first reported by The Guardian, the newly-instated robot editors of MSN.com selected a story about Little Mix singer Jade Thirlwall's experience with racism to appear on the homepage, but used a picture of Thirlwall's bandmate Leigh-Anne Pinnock to illustrate it. Thirlwall drew attention to the mistake on her Instagram story, writing: "@MSN If you're going to copy and paste articles from other accurate media outlets, you might want to make sure you're using an image of the correct mixed race member of the group." She added: "This shit happens to @leighannepinnock and I ALL THE TIME that it's become a running joke ... It offends me that you couldn't differentiate the two women of colour out of four members of a group ... DO BETTER!"
This discussion has been archived. No new comments can be posted.

Microsoft's AI Journalists Confuse Mixed-Race Little Mix Singers on MSN Homepage

Comments Filter:
  • by Anonymous Coward

    a turd of a story.

    • We cover website news inaccuracy around here.

      Oh, you were calling the MSN article the turd... my mistake.

      • The two women look very much alike and they admit that humans mix them up "all the time".

        So an "AI" made a human-like mistake.

        Does this mean it passes the Turing Test?

        • Microsoft can't be blamed for this, they have always said that using A.I. would be disastrous. I can now see why.
  • The AI software had simply chose the wrong picture to show next to the story.
    • by Anonymous Coward on Tuesday June 09, 2020 @11:49AM (#60164070)
      It involved a person of color, so the AI is racist. That's what passes for logic these days
      • by jandoe ( 6400032 )

        If it doesn't see the race then it is perfectly not racist, right?

      • by Anonymous Coward
        It involved calling an algorithm "AI". That's what passes for logic these days.
    • Why use an AI that makes such a mistake, clearly it isn't up to production level yet.

      I know a bunch of White Brogrammers made the code, and they never realized to code traits that make people look different.

    • Exactly.

      Now, the universe has changed IRT news photos, I'm sure. But way back when, one thing that surprised me was that reporters and photographers were separate. While they might send a reporter and a photographer to cover a story, there wasn't a necessarily a link between the two. You might have a caption that you could use with a text search (eg, "Find news stories whose text is like the caption and whose date matches the date of the news story") but that was about it. It was mostly intended to be d

    • by eepok ( 545733 )

      But I can't get outraged at a reasonable error... I don't like this explanation!

    • The AI doesn't create the articles, it just grabs them from other sources and posts. A human editor would more likely be able to spot that there was an error, especially if it's an editor that has experience in the subject matter. The title is misleading, and the summary implies that the AI made the original mistake, so it's another great day in the history of clickbaiting.

  • by gnasher719 ( 869701 ) on Tuesday June 09, 2020 @11:44AM (#60164038)
    The City of Birmingham, UK, printed 600,000 brochures telling people how nice their town was. Unfortunately they used stock photos of Birmingham, USA. All the brochures were destroyed.

    Doesn't excuse this reporting though.
    • The kind of people who wash up in local councils are one step up from unemployable. They're so useless and/or lazy no private company will touch them so the public sector is where they end up.

  • What has this got to do with race?
    • Nothing. It's about an AI and RGB values but someone got offended and decided to make it all about race for attention-seeking purposes.

      • Colour may have had little to do with it. Repeatedly appearing under the same banner in the same band might have done.
    • They were non-robotic so the AI stereotyped them as "generic human" and decided "all humans look the same" and substituted photo of one for another.

  • The end of racism (Score:4, Insightful)

    by Solandri ( 704621 ) on Tuesday June 09, 2020 @11:48AM (#60164058)
    The end of racism is not when we treat all races with equal respect. That implies that for whatever reason we're still keeping track of our differences, putting people into pigeonholes. (And unfairly crucifying anyone who unknowingly puts someone into the wrong pigeonhole.)

    The end of racism is when we don't care about race anymore.
    • The end of racism is not when we treat all races with equal respect. That implies that for whatever reason we're still keeping track of our differences, putting people into pigeonholes. (And unfairly crucifying anyone who unknowingly puts someone into the wrong pigeonhole.)

      The end of racism is when we don't care about race anymore.

      I view it somewhat differently. We should care about race, as well as other individually identifying characteristics, because these facets contribute to form part of our identities. Being blind to them erases a part of our identities.

      The problem is not acknowledging the existence of distinguishing characteristics (this part of the diversity that we desire) but addressing the inequalities in behavior that arise from prejudicial consideration of those characteristics. The problem is prejudice (e.g., I know

  • Hahaha, no that’s not a racist AI, if you want an example of that look no further than tay. I’m not sure what Microsoft was thinking by forcing an AI to gaze unblinking into the unfathomable human hell that is Twitter for its training dataset, but the outcome was obvious and inevitable.
  • When I google either of them images of both are mixed up along with a couple other people. In either case the most distinctive feature is the eyes and they are identical. Other than that one has a softer chin than the other but not so much that you can identify it clearly in all pictures. There are pictures where I can't clearly identify and I'm sure it is true of the journalists as well.

    The problem isn't that the AI can't distinguish, the problem is that the AI is relying on visual recognition to label and

  • No, you're a celebrity. You exist to entertain. Nobody cares how you feel.

  • by Murdoch5 ( 1563847 ) on Tuesday June 09, 2020 @01:10PM (#60164608) Homepage
    Deal with it, we'll correct the very minor issue, sorry your snowflake ass got offended, but our guess is you'd be offended even if the story / picture was accurate.
  • By any chance is the AI called Syncopatico? /probably obscure //w1a

  • by satanicat ( 239025 ) on Tuesday June 09, 2020 @02:28PM (#60165054)

    The messed up thing about this is that journalists are being laid off in favor of a software program. I know journalists have been in a bad way in recent years, but it's a really scary concept to think that eventually the only news we will get will be word of mouth, whatever X-social-media-platform decides to throw into your feed and from your government, and probably "intelligently fed to you" through an AI software.

    Sooner or later most stories we read will be cited with "cause I really think so"...

    I don't even really care about the picture thing, I mean they messed up... move along? I guess it does illustrate part of my point though, a proper journalist would probably do all the fact checking before posting...

    • In the future Microsoft will plug the revived tay chat bot into the comments section of their computer written news articles, allowing the comments section to be prefilled will racial slurs just like a real news site. This will close the loop and remove humans from the news business altogether.

    • Sooner or later most stories we read will be cited with "cause I really think so"...

      That's orthogonal to them being prepared by AI.

      • a proper journalist would probably do all the fact checking before posting

        Oh jeeze that's rich. Good thing we can't point to any instances of journalists getting facts wrong.

        • I think it's ok to be wrong. Provided that is transperant.

          I know what you mean, but there are ethical journalists out there.

          I wouldnt mind if Fox took on the AI though:)

  • the automated system confused two mixed-race members of British pop group Little Mix

    OH GOD no!!!

  • Really? Racist? (Score:5, Insightful)

    by mpercy ( 1085347 ) on Tuesday June 09, 2020 @03:41PM (#60165384)

    If I showed you a picture of ABBA, can you tell me which of the two (very) white women was Agnetha Fältskog and which was Anni-Frid Lyngstad? Or which of the men was Björn Ulvaeus and which was Benny Andersson? Especially if you went on Google to find pictures and some of them were wrong, but you'd made a good-faith effort to get it right and can show your sources?

  • Names of singers in a little-known British pop band gets mixed up by news outlet. I'm sure this has never happened before. And if it has, it has surely only happened to non-white people. Certainly, never in the history of journalism has a picture of a white person been accompanied by text mistakenly identifying that person as someone else. Hence, racism.
  • Comment removed based on user account deletion
  • by nagora ( 177841 ) on Tuesday June 09, 2020 @05:04PM (#60165796)

    They're all interchangeable non-entities. I imagine they struggle to tell themselves apart.

"Remember, extremism in the nondefense of moderation is not a virtue." -- Peter Neumann, about usenet

Working...