Comment on OpenAI confirms that AI writing detectors don’t work

• bioemerl@kbin.social ⁨1⁩ ⁨year⁩ ago

  Because you're training a detector on something that is designed to emulate regular languages closest possible, and human speech has so much incredible variability that it's almost impossible to identify if someone or something has been written by an AI.

  You can detect maybe your typical generic chat GPT type outputs, but you can characterize a conversation with chat GPT or any of the other much better local models (privacy and control are aspects which make them better) and after doing that you can get radically human seeming outputs that are totally different from anything chat GPT will output.

  In short, given a static block of text it's going to be nearly impossible to detect if it's coming from an AI. It's just too difficult to problem, and if you're going to solve it it's going to be immediately obsolete the next time someone fine tunes their own model

  source

  • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

    Yeah this makes a lot of sense considering the vastness of language and it’s imperfections (English I’m mostly looking at you, ya inbred fuck)

    Are there any other detection techniques that you know of? Wb forcing AI models to have a signature that is guaranteed to be indentifiable, permanent, and unique for each tuning produced? It’d have to be not directly noticeable but easy to calculate in order to prevent any “distractions” for the users.

    source

    • Grimy@lemmy.world ⁨1⁩ ⁨year⁩ ago

      The output is pure text so you would have to hide the signature in the response itself. On top of being useless since most users slightly modify the text after receiving it, it would probably have a negative effect on the quality. It’s also insanely complicated to train that kind of behavior into an llm.

      source

      • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

        Your implementation of my concept might be useless, but that doesn’t mean the concept is.

        One possible solution would be to look at how responses are structured, letter frequencies, etc. The flexibility/ambiguous nature natural language is that you can word things in many many different ways which allows for some creative meta techniques to accomplish a fingerprint.

        source

        • -> View More Comments
    • bioemerl@kbin.social ⁨1⁩ ⁨year⁩ ago

      forcing AI models to have a signature that is guaranteed to be indentifiable, permanent, and unique for each tuning produced

      Either AI remains entirely in the hands of fucks like open AI or this is impossible and easily removed. AI should be a free common use tool, not an extension of corporate control.

      source

      • roguetrick@kbin.social ⁨1⁩ ⁨year⁩ ago

        Owning the means of AI production huh? I guess anarchists will win after all.

        source

        • -> View More Comments
      • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

        Agreed, such power should belong to everyone or has yet to be discovered. Even Oppenheimer knew, once the cats out of the bag…

        source
• Eufalconimorph@discuss.tchncs.de ⁨1⁩ ⁨year⁩ ago

  Because AIs are (partly) trained by making AI detectors. If an AI can be distinguished from a natural intelligence, it’s not good enough at emulating intelligence. If an AI detector can reliably distinguish AI from humans, the AI companies will use that detector to train their next AI.

  source

  • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

    I’m not sure I’m following your argument here - you keep switching between talking about AI and AI detectors. Each of the below are just numbered according to the order of your prior responses as sentences:

    1. Can you provide any articles or blog posts from AI companies for this or point me in the right direction?
    2. Agreed
    3. Right…

    I’m having trouble finding your support for your claim

    source

    • Theharpyeagle@lemmy.world ⁨1⁩ ⁨year⁩ ago

      See Generative Adversarial Network (GAN). Basically, making new AI detectors will always be harder than beating current ones. AI detectors have to somehow find a new “tell”, the target AI need only train itself on the output of the detector to figure out how to trick it.

      source

      • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

        ChatGPT isn’t a GAN network.

        source
    • dack@lemmy.world ⁨1⁩ ⁨year⁩ ago

      At a very high level, training is something like:

      • generate some output
      • give the output a score based on how much it looks like real human text
      • adjust the parameters slightly to improve the score
      • repeat

      Step #2 is also exactly what an “AI detector” does. If someone is able to write code that reliably distinguishes between AI and human text, then AI developers would plug it in to that training step in order to improve their AI.

      In other words, if some theoretical machine perfectly “knows” the difference between generated and human text, then the same machine can also be used to make text that is indistinguishable from human text.

      source

      • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

        Exactly right, I mentioned this in a comment elsewhere but basically we can’t have our cake and eat it too.

        We can’t have a perfect NL impersonator that can also be detected as not NL. (Best case, obviously things are perfect for any AI model so technically detecting those mistakes could be used to help identify perhaps, but who’s to say what the FP rate would look like!)

        Ultimately the cat is out of the bag and I’m not quite sure there is anything we can do now. Ultimately some smart fingerprinting solution would be ideal but I just don’t know how feasible that would remain.

        source
• sebi@lemmy.world ⁨1⁩ ⁨year⁩ ago

  Because generative Neural Networks always have some random noise. Read more about it here

  source

  • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

    Isn’t that article about GANs?

    Isn’t GPT not a GAN?

    source

    • PetDinosaurs@lemmy.world ⁨1⁩ ⁨year⁩ ago

      It almost certainly has some gan-like pieces.

      Gans are part of the NN toolbox, like cnns and rnns and such.

      Basically all commercial algorithms (not just nns, everything) are what I like to call “hybrid” methods, which means keep throwing different tools at it until things work well enough.

      source

      • stevedidWHAT@lemmy.world ⁨1⁩ ⁨year⁩ ago

        The findings were for GAN models, not GAN like components though. That’s bad science right there.

        source

        • -> View More Comments
    • bioemerl@kbin.social ⁨1⁩ ⁨year⁩ ago

      It's not even about diffusion models. Adversarial networks are basically obsolete

      source