cross-posted from: https://beehaw.org/post/7458945

Many mushroom identification and foraging books being sold on Amazon are likely generated by AI with no human authorship. These books could provide dangerous misinformation and potentially lead to deaths if people eat poisonous mushrooms based on the AI’s inaccurate descriptions. Two New York mushroom societies have warned about the risks of AI-generated foraging guides. Experts note that safely identifying wild mushrooms requires careful research and experience that an AI system does not have. Amazon has since removed some books flagged as AI-generated, but more may exist. Detecting AI-generated books and authors can be difficult as the systems can fabricate author bios and images. Relying on multiple credible sources, as well as guidance from local foraging groups, is advised for safely pursuing mushroom foraging.

    • alternative_factor
      link
      fedilink
      1810 months ago

      That’s wild that this article popped up on my feed this instant, I saw an ad for AI plant identification through your camera lens. The first thought I had was “holy shit so many people are going to die when Alexa misidentifies water hemlock as cow parsnip”.

      • @d3m0nr4v3r@feddit.deM
        link
        fedilink
        210 months ago

        Tbf the same thing exists for mushrooms, there are various apps. It’s dangerous to rely on them but I use them for a first idea of what I might be looking at and then go from there. Quite useful sometimes

        • alternative_factor
          link
          fedilink
          110 months ago

          Yeah I have nothing against the idea itself, I just know there are people out there who will trust an app 100% and never try to go from there. It’s super sad really, if some people had just a little more discernment I wouldn’t instantly have thought about people dying from it.

    • @warbond@lemmy.world
      link
      fedilink
      410 months ago

      Holy shit, I hadn’t even thought of that. Does Amazon have a responsibility to ensure the information it sells is at least, “commonly thought to be true”(or whatever)? What would the metric be? Who could claim jurisprudence?

      IANAL, but I love the implications of case law.

      • @Shdwdrgn@mander.xyz
        link
        fedilink
        English
        810 months ago

        Probably not because they can afford good lawyers, but they SHOULD bear some responsibility. I think if someone actually dies because of this we might see the limits of the laws tested.

  • VardøgorM
    link
    fedilink
    2510 months ago

    i never thought about physical books getting AI generated! that’s scary.

    • MuchPineapples
      link
      fedilink
      810 months ago

      They sometimes also claim to be written by (semi-)famous authors, so it’s also fraud.

    • hallettj
      link
      fedilink
      English
      310 months ago

      I think these are ebooks. You can self-publish online for free so there isn’t much of a barrier to putting something out there, and hoping a few people fall for it and buy copies.

      • Square Singer
        link
        fedilink
        810 months ago

        Not only ebooks. Amazon offers a print-on-demand service where you have basically no barrier to entry. You upload your book in digital form and if someone orders it, they print a copy and send it.

        There’s a huge amount of trash on that service.

        • hallettj
          link
          fedilink
          English
          110 months ago

          Oh interesting! That sounds like a nifty service if you set aside the trash.

    • _cnt0
      link
      fedilink
      510 months ago

      By “they”, do you mean AIs, or capitalists? There currently is no artificial intelligence, only things that are called that for marketing purposes. Those things are used by capitalists to, shocker one: accumulate capital, and, shocker two: capitalists don’t have a history of caring about the general wellbeing.