Some of the most harmful applications of AI tools are not hidden on the dark corners of the internet, but are actively promoted to users by social media companies.

Instagram is profiting from several ads that invite people to create nonconsensual nude images with AI image generation apps, once again showing that some of the most harmful applications of AI tools are not hidden on the dark corners of the internet, but are actively promoted to users by social media companies unable or unwilling to enforce their policies about who can buy ads on their platforms.

While parent company Meta’s Ad Library, which archives ads on its platforms, who paid for them, and where and when they were posted, shows that the company has taken down several of these ads previously, many ads that explicitly invited users to create nudes and some ad buyers were up until I reached out to Meta for comment. Some of these ads were for the best known nonconsensual “undress” or “nudify” services on the internet.

    • kbin_space_program@kbin.run
      212·
      2 years ago

      Nice pipe dream, but the current fundamental model of AI is not and cannot be made deterministic. Until that fundamental chamge is developed, it isnt possible.

      • fuckthepolice@lemm.eeEnglish
        13·
        2 years ago

        the current fundamental model of AI is not and cannot be made deterministic.

        I have to constantly remind people about this very simple fact of AI modeling right now. Keep up the good work!

            • kbin_space_program@kbin.run
              42·
              2 years ago

              I love your wishful thinking. Too bad academia doesnt agree with you.

              Edit: also, I have to come back to laugh at you for trying to argue that the almost random nature of software random number generators is deterministic AI.

                • kbin_space_program@kbin.run
                  42·
                  2 years ago

                  You can make it more deterministic by reducing the acceptable range of answers, absolutely. But then you also limit your output, so thats never really a good use case.

                  Randomness is a core functionality of not just LLMs, but the entire stack that has resulted in LLMs. Yes you can get a decently consistent answer, but not a deterministic one. Put another way, with LLMs being at max constraint, you can ask them to add 1+1. You’ll usually get 2. But not nearly always.