The closures, so employees can watch a recorded lecture, will cost the company an estimated 2.1bn won ($1.4m) in sales

That AI is at fault here, is hidden as an aside in the article.

Sure, the humans should have been better in all this, but that AI even proposed such thing on that date, makes it seem ultra evil. If AI would be able to think or have a conscience, that is.

  • FriendOfDeSoto@startrek.websiteEnglish
    7·
    7 hours ago

    Starbucks said it was “deeply sorry for an unacceptable marketing incident”

    An unacceptable marketing incident? Like it was thrust upon them. Incidentally, by themselves. The language is baffling although I’m willing to accept it might have sounded less incidental in the original Korean.

    Obvs this is very dumb, a bunch of executives are asleep at the wheel, and they probably deserve the 26% drop in revenue for it. The apology needs to be performative and this company-wide, revenue sacrificing history lesson is coincidentally good marketing, I imagine.

    That being said, this isn’t really a FuckAI story. Sure, the model suggested a couple of things that only sound great if you’re ignorant of Korean history. But this isn’t a model breaking the sandbox and releasing all company secrets by accident. Or suggesting to help teenagers with ending their lives. This is first and foremost careless managerial conduct. They should have stopped the campaign and even might have if they had bothered to read the email attachments. This could’ve come from a historically ignorant ad agency as well. This is human error, not model madness.

    Looking forward to the US Starbucks smashing caffeine explosion campaign that will surely bring the house down in September.

    • kamenladyOP
      2·
      6 hours ago

      this isn’t a model breaking the sandbox and releasing all company secrets by accident. Or suggesting to help teenagers with ending their lives. This is first and foremost careless managerial conduct.

      I agree with this. My emphasis was on it pointing out things that only could have been “great marketing”, if it wasn’t for it being the things that happened on that exact day. I got a little carried away with it

      I am curious about what the exact prompt was though. Depending on how unfortunate the prompt was formulated, the suggestion being all about the incident could be very much a valid response.

  • hexagonwin@lemmy.today
    4·
    6 hours ago

    bro this isn’t related to ai at all. as a korean i can only say this was 100% deliberate and intentional, it only could be ai output if the input explicitly mentioned their intention to mock the gwangju democratic uprising (and others)

    i’m confused why this is even posted in the fuck ai comm…

  • Daniel Quinn@lemmy.caEnglish
    4·
    6 hours ago

    We have to be more deliberate with our language . the “AI” isn’t at fault because these things can’t be be held accountable. The marketing executives that chose to trust sparkly autocomplete instead of doing their jobs are at fault, and one might even extend the blame to include the companies that trained the model and provided the service.