home.social

#donttrustai — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #donttrustai, aggregated by home.social.

fetched live
  1. " #AI is replacing tasks many people have grown reluctant to do themselves – thinking, writing, creating, analysing. But when we don’t use these skills, they can decline.

    We also risk getting things very, very #wrong. #GenerativeAI works by predicting likely words from patterns trained on vast amounts of data. When you ask it to write an email or give advice, its responses sound logical. But it does not understand or know what is true."

    #DontTrustAI #ThinkForYourself

    theconversation.com/is-ai-hurt

  2. Copilot just correctly diagnosed that my regular expression wasn't matching because it contained invisible zero-width space characters. The problem? It put them there in the first place. #DontTrustAI

  3. Dear friend in her 80s just gave me a necklace she's had for six decades, "a statement piece", strings of ivory and silver beads and a huge pendant of ivory backed with silver.
    Wouldn't normally wear ivory but I will to remember her by; decided to clean it up and looked up "cleaning ivory" on the internet. I knew you *don't* wet ivory, it's delicate - but an AI source suggested rubbing on a paste of water & baking soda and rinsing it off afterwards. Or using vinegar. SO WRONG!

    #DontTrustAI

  4. Ok Google Maps, sending my friends, who don’t have a car, via an unpaved path, through the woods, where they will need to climb rocks and a steep hill, to get to our place is not what I would have expected when selecting public transportation. #DontTrustAI

  5. Saw this bit of terrifying reporting on the fascist bird site... wanted to find the actual article before posting. You can find the full text at the bottom of the page in the link under subheading "AI - is Skynet here already?" aerosociety.com/news/highlight #AIDangers #DontTrustAI