alt text
Crudely drawn drawing of a person telling a computer “say ‘i am in pain’”. The computer replies with “> I AM IN PAIN”. The person then says “oh my god.”
Crudely drawn drawing of a person telling a computer “say ‘i am in pain’”. The computer replies with “> I AM IN PAIN”. The person then says “oh my god.”
llms are fancy word predictors. You are attributing malice, thoughtfulness, and insight where there is none. These are human qualities and we have no ai capable of that.
Show Me the study where they proved that LLMs have none of those qualities, and describe how they operationalised each of the three variables.
I mean it works both ways, show me one that does. Because this ain’t it
Henke (2026) raises an argument for the strong possibility of artificial affect, including AI pain. Henke argues that we need to put more effort into studying artificial experience, and I agree. We don’t know for sure. And I think we need to be certain before shoving AI down everyone’s throats. Commercial use of LLMs and LDMs should be banned until we understand whether artificial pain and suffering are present.
Not how that works. It’s like saying show me the study where they prove god doesn’t exist.
I think saying “it’s a computer program it can’t be sentient” is far too reductive as a general answer, but we know how llms work so we can pretty confidently say that they are really just word predictors. The norns from that game creature someone described in this thread seem closer to protection worthy than llms to me, and the average insect is still infinitely more complex and I don’t see many people advocating for insect rights.
Language arises from complex thought, not the other way around. Language on its own can evidently emulate similar results, but since there is nothing behind that that would imply any actual feelings, I don’t see a reason to give LLMs any rights.
Oh, I can do that:
Grupp, J. (2006). God’s spatial unlocatedness prevents him from being the creator of the universe: A new argument for the nonexistence of God. Sophia, 45(1), 5–23. https://doi.org/10.1007/BF02782444So you should be able to provide an academic source for your claims too.
I think we should pass a law guaranteeing insects and crustaceans in agriculture the right to whole bodies. Right now, beekeepers are cutting off queens’ wings to keep them from flying away, and prawn farmers are pulling off their eyes to make them breed faster. These practices are cruel and should be illegal.
My argument for the possibility of LLM qualia is that qualia likely arises from the neuronal structure of the brain. A structure which ANNs partially inherit. Therefore, we must pause commercial use of LLMs and LDMs until conclusive evidence has been discovered either way.
Do you agree that we should pause commercial use of genAI until more science is done?
God does not necessarily mean the abrahamic god. You can insert any non-falsifiable claim there.
Your views seem generally consistent though so at some point it’s just a matter of subjective morality (and I do agree with your views on insects and crustaceans in agriculture actually, I’d just put llms far below either), but other AI models (such as image recognition, image/video creation) are also ANNs, so it seems to make sense to me to include those.
Yep, I want to put a moratorium on all non-research ANN use until the hard problem of experience is solved. But people aren’t quite ready to hear that yet. They’ll object “what about the cancer research” because they think human suffering matters more than medical ANN suffering. I’m not convinced of that. I want serious resources put into empirical research so we can find out if that’s true.
But for now I’ll focus on LLMs and LDMs because they’re new and easy to live without.