The backlash was immediate, but it didn’t stop the BBC from using text generated by LLMs—and purportedly checked and copy-edited by a human before approval—in two marketing emails and mobile push notifications to advertise Doctor Who. But now, the corporation will stop the experimentation entirely after a wave of official complaints pushed them to offer a response to concerned audiences.

  • Kusimulkku@lemm.ee
    link
    fedilink
    English
    arrow-up
    27
    arrow-down
    3
    ·
    7 months ago

    BBC from using text generated by LLMs—and purportedly checked and copy-edited by a human before approval—in two marketing emails and mobile push notifications to advertise Doctor Who.

    If they’re telling the truth then I don’t really get what’s wrong about that particular use

    • yogurt@lemm.ee
      link
      fedilink
      English
      arrow-up
      7
      ·
      7 months ago

      They weren’t just being cheap, they wanted to take human-written ads, auto-generate a million variations to send to individual people, then feed the ones that got clicks back in to train an AI clickbait generator. It also means the variations would be functionally watermarked so if anybody posted part of their text on Reddit or something the BBC could track who they sent that variation to.