• 10 Posts
  • 145 Comments
Joined 2 years ago
cake
Cake day: July 24th, 2023

help-circle
















  • We’re talking about home use AI searches… you said it was unethical so maybe you should define exactly why you think this?

    Today I wanted to know what the tyre pressures should be for my 2002 Corolla and AI gave me the answer, I would not have bought a book or gone anywhere past the first page of google for that information.

    The possible ethical dilemma is depriving someone of compensation because I used their research and deprived them of potential revenue, in reality I would never have bought a book on tyre pressures or car maintenance, and it’s unlikely I would ever have visited a site where adverts would have paid the contributors.

    Another dilemma is of power consumption, the model is already made then it’s already used the power, and my tiny LLM query is going to use far less power locally than a web based search.

    As a company who might make money, or achieve cost savings from using AI trained on data some only intended for use by a human, I can see how this is not always ethical.


    1. Whose data is it trained on? Seems like an ethical dilemma to me.

    Using a standalone LLM for personal use doesn’t seem like an ethical dilemma to me, it’s already been trained on the data and if the data was accessible on the web or via a library then I don’t see the harm.

    Getting small amounts of medium-trust information on a subject, is a good way to get someone interested enough to read a book, watcha a YouTube video or find a website for more information and validate the AI response.