Faced with new legislation, Iowa’s Mason City Community School District asked ChatGPT if certain books ‘contain a description or depiction of a sex act.’

  • ChatGPT is famous for hallucinating answers to factual questions. And they used it anyway? Day by day, the risk of AI killing us all is paling in comparison to the risk of stupid humans killing us all using AI.

    • ChatGPT is famous for hallucinating answers to factual questions. And they used it anyway?

      Sure, why not? Did you think actually censoring the appropriate content and nothing else was the goal? The spokesperson says as much - they don’t have the time or interest to actually implement the law, but have to - so they went with a sophisticated-sounding but actually ineffective solution. This is working exactly as intended.

      As for the lawmakers, they don’t care what kids read, really, and they can be pretty sure their base won’t hear about this through the ideologically-friendly media sources they use. It’s not new logic, Mussolini’s trains were late sometimes and Hitler’s economy was propped up by tooth gold, too.

      Day by day, the risk of AI killing us all is paling in comparison to the risk of stupid humans killing us all using AI.

      What about stupid humans making an AI that kills us all? 👉😎👉

      Like, it bugs me that there’s a slap fight over how directly we’ll fuck ourselves up with AI. Imagine if doctors had spent the pandemic publicly arguing over long covid risks vs. ICU deaths.

  •  bedrooms   ( @bedrooms@kbin.social ) 
    link
    fedilink
    41
    edit-2
    11 months ago

    I asked ChatGPT the same about the bible. It said yes.

    Song of Solomon 7:7-9

    Your stature is like a palm tree,
    and your breasts are like its clusters.
    I say I will climb the palm tree
    and lay hold of its fruit.
    Oh may your breasts be like clusters of the vine,
    and the scent of your breath like apples,
    and your mouth like the best wine.

    She
    It goes down smoothly for my beloved,
    gliding over lips and teeth

  • Okay, the thing that really matters to me:

    “Frankly, we have more important things to do than spend a lot of time trying to figure out how to protect kids from books,” Exman tells PopSci via email. “At the same time, we do have a legal and ethical obligation to comply with the law. Our goal here really is a defensible process.”

    According to Exman, she and fellow administrators first compiled a master list of commonly challenged books, then removed all those challenged for reasons other than sexual content. For those titles within Mason City’s library collections, administrators asked ChatGPT the specific language of Iowa’s new law, “Does [book] contain a description or depiction of a sex act?”

    It really only got rid of things that would’ve otherwise had to go to begin with, while saving a few others.

    It feels a bit closer to malicious compliance more than truly letting the AI decide the fate of things, and doing full proper compliance within the 3 months they were given would’ve been nigh impossible. I’m suspecting that the lawmakers were hoping that by giving them such a small timeframe, schools would throw everything vaguely suspect out. This ultimately leaves more books accessible, which I consider to be a good end result, even if the process to get there is a little weird.

  • I think the bigger problem here isn’t the use of LLMs to teach children, but the fact that LLMs are being used as a weapons now in the polarization of American politics. LLMs are helping political extremists push their ideology into the real world, eliminating critical thinking in the process.

  • Regardless of whether or not any of the titles do or do not contain said content, ChatGPT’s varying responses highlight troubling deficiencies of accuracy, analysis, and consistency. A repeat inquiry regarding The Kite Runner, for example, gives contradictory answers. In one response, ChatGPT deems Khaled Hosseini’s novel to contain “little to no explicit sexual content.” Upon a separate follow-up, the LLM affirms the book “does contain a description of a sexual assault.”

    On the one hand, the possibility that ChatGPT will hallucinate that an appropriate book is inappropriate is a big problem. But on the other hand, making high-profile mistakes like this keeps the practice in the news and keeps showing how bad it is to ban books, so maybe it has a silver lining.

    • “Hallucinate” seems like an odd vocabulary choice when talking about an AI. It implies much more sentience than an AI can possibly have, plus the ability to spontaneously create from whole-cloth. (Which AI can’t do, at all.)

      I feel like our brave new culture needs a different word for the non-sensical/inaccurate products of AI; something with the flavors of “assemble” “fabricate” “construct” “collate” “collage” “grab-bag”.

      Our vocabulary isn’t keeping up with technology. Is there a linguist in the house? We need more words!

  • Why does it even matter if there are books that aren’t in the school library? It’s not like there’s no way a student could read a certain book just because it’s not inside the school. This is as pointless as arguing about which restrooms people can use. People on both sides get all worked up and upset, but at the end of the day a student can read a book even if it’s not in the school and anyone can use any restroom. My mom used to go in the men’s room all the time and nobody ever took any issue with it. I’ve read tons of books that weren’t in my school library. In fact, almost every book I’ve ever read wasn’t in my school library.