12 Matching Annotations
  1. Oct 2025
    1. But the pages of e-books are themselves likely to become the sites of conversations. Already readers ofmany e-books-on the Kindle, the Nook, and other e-readersshare comments and highlight

      this is kind of what we are doing now with Hypothesis.

  2. Sep 2025
    1. “If these new technologies only speak Western languages, we’re now excluded from the digitaleconomy,” says Running Wolf. “And if you can’t even function in the digital economy, it’s going to bereally hard for [our languages] to thrive.”

      New technologies remove access to many people by only using certain languages. this could be a contributor to the eradication or fading of certain languages if these new technologies do not begin to represent more languages.

    1. “White supremacist and misogynistic, ageist, etc., views are overrepresented inthe training data, not only exceeding their prevalence in the general population but alsosetting up models trained on these datasets to further amplify biases and harms.”

      this reminds me of a previous article where the authors mentioned how AI trains itself on human history, including the harmful stereotypes. As a result, AI has been trained on these harmful things and regurgitates that rhetoric.

    1. Furthermore, AI models are prone to biases, errors, and adversarial vulnerabilities,

      I am interested in researching this more as I am curious about how to prevent AI from perpetuating harmful stereotypes and pushing out harmful rhetoric.

    2. I edited the sentence to sound less like AI writing.

      I actually do this very often and it bothers me that there is becoming less of a distinction between someone with good grammar and rhetorical techniques and someone using AI.

    1. We should notworry about whether the product of their work is economically valuable, or whether it could be createdby more efficient mean

      This efficiency issue has become very ingrained within our society, people are constantly asking themselves how they can make things easier or quicker without caring so much about how they get there. reminds me of the phrase its not about the journey its about the destination. I think the journey is just as important as the destination.

    1. I use AI a lot. Like, every day,” she said. “And I do believe it could take awaythat critical-thinking part. But it’s just — now that we rely on it, we can’t really imagine living without it

      It is interesting how we are now growing up in an age where technology is totally embedded within our society. People have begun to use it as a crutch, something you depend on. I think this sort of mentality can be dangerous as it does remove the necessary analytical aspect of learning.

    1. Work where you need a specific perspective, and where a simulated first pass from that perspective can be helpful, like reactions from fictional personas.

      This can be a beneficial way to use AI as you are using it as a tool, not a crutch, not a substitute for true human analysis. I think this way of using AI is a good way to use it.

    1. which are trained on vast swathes of humanity’s cultural heritage, can often best be wielded by people who have a knowledge of that heritage

      Essentially, because AI trains itself on human's history, it is best understood and used by people who also have extensive knowledge of human history. This is because they could fact check or ask it the right questions.

    1. hey also need to know that some products, including large-language model text generators, have clear harms and dangers now

      This ties into a larger issue of how AI can perpetuate harmful stereotypes within its responses.