Close Menu
    Technik

    Google entschuldigt sich, nachdem AI News Alert über das Debakel bei den BAFTA Film Awards das N-Wort enthielt

    Share.
    View 14 Comments

    14 Kommentare

    1. ddgconsultant on

      This is a clear example of why AI-generated content summaries need more robust review mechanisms before reaching users. Google’s AI likely surfaced the term from the article it was summarizing, but without proper content filtering, it generated a harmful output. The bigger concern is the speed at which AI news alerts propagate – by the time a correction is issued, the damage is done. Responsible AI deployment in high-visibility features like news summaries requires multilayer safety checks, not just post-incident apologies.

    4. justbunnies on

      I saw the screen shot and I’m shocked how blatant it is. This is what happens when you use the internet and social media to teach your AI.

    5. jpiro on

      First case of an AI agent suffering from Tourette’s Syndrome. They really are becoming more like us every day.

      /s

    7. Night-Monkey15 on

      I thought it’d be at least in a quote or something. Nope. This is just blatant racism

    9. into_wishin_666 on

      People keep citing what needs to be done to A.I., except for the one true solution, scrapping it.

    10. Dreddddddd on

      Google, I don’t care how many passes you trained your model on, you don’t have the pass.

    11. Elliot-S9 on

      Generative AI is so useless. Can we please just admit this and go back to 2016? 

    12. CharacterCompany7224 on

      Okay any consequences or just another apology also written by AI?

    14. knightcrusader on

      They since added this „clarification“:

      > The original version of this story stated that the Google news alert was AI-generated. Google has since made clear that, although the inclusion of the N-word was a tech error, it was not AI-generated. Google said its systems “recognised a euphemism for an offensive term on several web pages, and accidentally applied the offensive term to the notification text.” It added: “This system error did not involve AI. Our safety filters did not properly trigger, which is what caused this.“

      Um, so the system did it on its own without human oversight? Picking words based on its usage in other places? Hmm, sounds like something an LLM would do…

    Leave A Reply