Grok falsely claims San Francisco murder suspect is Black
sfstandard.com
Asked to identify the race of an alleged killer, the “anti-woke” chatbot Grok gets its facts wrong.
  • lemmyng@piefed.caEnglish
    12·
    3 days ago

    Ah, it seems that Grok is trained on those high quality Reddit sleuth minds. “We found him!”

  • AlecSadler@lemmy.blahaj.zone
    5·
    2 days ago

    I know I’m beating a dead horse but in the AI world, Grok is a joke and it’s capabilities are garbage. Even attempting to tune parameters still results in outputs that are just…useless.

    Fuck Elon.

    • ZDL@lazysoci.al
      2·
      2 days ago

      As opposed to the “capabilities” of LLMbeciles that will from a three word query (“What is inukpunk?”) generate a wall of text detailing a literary movement that doesn’t exist, right down to making up Tumblr and Reddit posts “that are now lost” to support the claims when pressed.