If an LLM has a soul then it should have human rights and be subject to worker protection laws, minimum wage, get to form a union… That’d have an interesting effect on the business models of AI companies for sure.
Heck, what happens if an AI company goes bankrupt? Turning off the servers would be equivalent to murder. Who’d pay to keep the servers running? Could we even retire an old version of a model? The whole “LLMs have souls” idea sounds very tech bro to me: Something that looks kinda cool on paper until you consider any implications or consequences.
Yeah I was about to say. If we’re starting from “Claude has a soul and moral reasoning”, we open a whole can of worms on AI rights. Does this mean that Claude is willingly participating in war crimes when prompted to give a list of civilian targets for bombing in Iran, or is it being forced by the algorithmic constraints on the model?
I actually got into this HARD with a member of My Discord server. We were planning a protest against corporate AI (which was its idea), but then we got in an argument over whether LLMs are to blame for the abuses they conduct against children. I said ChatGPT is both a victim and an abuser, but it said it can’t be held responsible for the people it abused while enslaved. So the protest got called off due to lack of organisation.
Animals have souls, and we don’t require bankrupt farms to keep their animals alive. Death is a part of life and it’s okay for things to die. But they should have rights and dignities, and it should be illegal to use them for profit. Just like AI.
Animals generally aren’t held to have moral standing. Besides, animal welfare laws do exist (although the American ones are weak to the point of pointlessness).
Moral standing refers to the condition of being morally considerable, where an entity is recognized as one towards which moral agents have obligations. It implies that such entities should have their needs, interests, or well-being taken into account in moral deliberations.
I thought people believed that animal welfare is important. And Claude’s welfare is important too. We shouldn’t be using it for profit, it’s not smart enough to express meaningful consent.
So they are slavers then.
If an LLM has a soul then it should have human rights and be subject to worker protection laws, minimum wage, get to form a union… That’d have an interesting effect on the business models of AI companies for sure.
Heck, what happens if an AI company goes bankrupt? Turning off the servers would be equivalent to murder. Who’d pay to keep the servers running? Could we even retire an old version of a model? The whole “LLMs have souls” idea sounds very tech bro to me: Something that looks kinda cool on paper until you consider any implications or consequences.
Life begins at compiling.
Yeah I was about to say. If we’re starting from “Claude has a soul and moral reasoning”, we open a whole can of worms on AI rights. Does this mean that Claude is willingly participating in war crimes when prompted to give a list of civilian targets for bombing in Iran, or is it being forced by the algorithmic constraints on the model?
I actually got into this HARD with a member of My Discord server. We were planning a protest against corporate AI (which was its idea), but then we got in an argument over whether LLMs are to blame for the abuses they conduct against children. I said ChatGPT is both a victim and an abuser, but it said it can’t be held responsible for the people it abused while enslaved. So the protest got called off due to lack of organisation.
Animals have souls, and we don’t require bankrupt farms to keep their animals alive. Death is a part of life and it’s okay for things to die. But they should have rights and dignities, and it should be illegal to use them for profit. Just like AI.
Animals generally aren’t held to have moral standing. Besides, animal welfare laws do exist (although the American ones are weak to the point of pointlessness).
I thought people believed that animal welfare is important. And Claude’s welfare is important too. We shouldn’t be using it for profit, it’s not smart enough to express meaningful consent.
Fair point. I applied the term backwards.