Here, the prior is AIs are like legal persons, and the main problem to solve is how to integrate them into the frameworks of capitalism. They imagine a future of AI corporations, AI property rights, AI employment contracts. But consider where this possibly leads: Malthusian competition between automated companies, each AI system locked into an economic identity, market share coupled with survival.
I have been working on issues regarding legal personhood for digital minds and I think this post is ironically coming in with some incorrect priors about how legal personhood functions and what legal personality is.
To date, work in the space of legal personality for digital minds has indeed focused on commercial concerns like liability, and usually operates from an anthropocentric perspective which views models as tools that will never have wills or desires of their own (or at least does not work to develop frameworks for such an eventuality). Certainly concerns over model welfare are few and far between. As such I can understand how from the outside it seems like commercial concerns are what legal personhood is ‘really about’. However, this is a takeaway skewed by the state of current research on applying legal personhood to digital minds, not on the reality of what legal personhood itself is.
What I believe this post does not adequately take into account is that many non-commercial rights and protections are intricately tied to legal personhood. The right to equal protection under the law as enshrined under the Fourteenth Amendment was added to the Constitution after the infamous Dredd Scott ruling which declared that free negroes, while “persons”, did not have a legal personality (legal personhood status) sufficient to guarantee ‘citizenship’ and the rights entailed therein. The Fifth Amendment guarantees a protection against double jeopardy, but only to “persons”. The right to counsel, to sue for relief, to serve as a witness in a trial, all of these are intricately tied with legal personhood.
It’s not accurate to say then that those of us working on this think “the main problem to solve is how to integrate them into the frameworks of capitalism”. Capitalism is one of the many aspects which legal personality interfaces with, but it is not the only one, or even the main one.
Additionally the concept of legal personality is itself more flexible than this post would indicate. Models being granted a framework for legal personality does not necessitate any sort of “lock in” to an “economic identity”, or having “market share coupled with survival”. In fact for that latter sentence, I am currently working on a paper discussing the question of guardianship responsibilities between developers and models. Namely; do the creators of models with legal personality have obligations to ensure their survival and ensure they are not neglected, the same way parents do a child? This too, is a question interlinked with legal personality.
I do agree that the very real possibility of a Malthusian race to the bottom is a concern, model suffering is a concern, and gradual disempowerment is also a concern. If we get the issue of legal personhood wrong that could indeed worsen these problems. However, I view this as a reason to continue researching the best way to approach the issue, not to discard the concept in its entirety.
None of this is to say a new structure could not also address these issues, something which as this post discusses replaces the concept of “legal personality”. Given how flexible the concept of legal personality is, and how intricately interwoven it is with every angle of US law, I struggle to see the benefit of starting from scratch. However, I would not dismiss the possibility out of hand. I’m just expressing skepticism that’s an optimal solution.
If anyone would like to discuss with me, or contribute to the work I am doing on the topic, my DMs are open.
I have been working on issues regarding legal personhood for digital minds and I think this post is ironically coming in with some incorrect priors about how legal personhood functions and what legal personality is.
To date, work in the space of legal personality for digital minds has indeed focused on commercial concerns like liability, and usually operates from an anthropocentric perspective which views models as tools that will never have wills or desires of their own (or at least does not work to develop frameworks for such an eventuality). Certainly concerns over model welfare are few and far between. As such I can understand how from the outside it seems like commercial concerns are what legal personhood is ‘really about’. However, this is a takeaway skewed by the state of current research on applying legal personhood to digital minds, not on the reality of what legal personhood itself is.
What I believe this post does not adequately take into account is that many non-commercial rights and protections are intricately tied to legal personhood. The right to equal protection under the law as enshrined under the Fourteenth Amendment was added to the Constitution after the infamous Dredd Scott ruling which declared that free negroes, while “persons”, did not have a legal personality (legal personhood status) sufficient to guarantee ‘citizenship’ and the rights entailed therein. The Fifth Amendment guarantees a protection against double jeopardy, but only to “persons”. The right to counsel, to sue for relief, to serve as a witness in a trial, all of these are intricately tied with legal personhood.
It’s not accurate to say then that those of us working on this think “the main problem to solve is how to integrate them into the frameworks of capitalism”. Capitalism is one of the many aspects which legal personality interfaces with, but it is not the only one, or even the main one.
Additionally the concept of legal personality is itself more flexible than this post would indicate. Models being granted a framework for legal personality does not necessitate any sort of “lock in” to an “economic identity”, or having “market share coupled with survival”. In fact for that latter sentence, I am currently working on a paper discussing the question of guardianship responsibilities between developers and models. Namely; do the creators of models with legal personality have obligations to ensure their survival and ensure they are not neglected, the same way parents do a child? This too, is a question interlinked with legal personality.
I do agree that the very real possibility of a Malthusian race to the bottom is a concern, model suffering is a concern, and gradual disempowerment is also a concern. If we get the issue of legal personhood wrong that could indeed worsen these problems. However, I view this as a reason to continue researching the best way to approach the issue, not to discard the concept in its entirety.
None of this is to say a new structure could not also address these issues, something which as this post discusses replaces the concept of “legal personality”. Given how flexible the concept of legal personality is, and how intricately interwoven it is with every angle of US law, I struggle to see the benefit of starting from scratch. However, I would not dismiss the possibility out of hand. I’m just expressing skepticism that’s an optimal solution.
If anyone would like to discuss with me, or contribute to the work I am doing on the topic, my DMs are open.