human.json and therian identities #52

mosssystem commented 2026年04月03日 07:55:12 +02:00

I'm going to put it out here for perusal that not everyone considers their identity as human.
From therianthropy, to furries, to all varieties of alterhumanity, there's a lot of different ways one can find themselves not feeling comfortable adopting human as a label.

This is especially true with the goal of this standard being a very personal pledge; as there's not really an enforceable division, it makes maintaining a human.json file a personal choice.

I admit, this is a concern from a minority perspective and may not resonate with many readers. However, I'm pitching it my concern because the time to do so is early on in a protocol's formation than at the end.

Other words that might be good candidates:

• authorship
• creation
• craft
• person

I'm going to put it out here for perusal that not everyone considers their identity as human. From therianthropy, to furries, to all varieties of alterhumanity, there's a lot of different ways one can find themselves not feeling comfortable adopting `human` as a label. This is especially true with the goal of this standard being a very personal pledge; as there's not really an enforceable division, it makes maintaining a `human.json` file a personal choice. I admit, this is a concern from a minority perspective and may not resonate with many readers. However, I'm pitching it my concern because the time to do so is early on in a protocol's formation than at the end. Other words that might be good candidates: * authorship * creation * craft * person

addisoncrump commented 2026年04月04日 19:47:57 +02:00

FWIW, one can, as of right now, specify the human.json to any name of their choosing. If one does not wish to associate themselves accordingly, I think it makes perfect sense to use, e.g.: <link rel="human-json" href="/being.json">.

The problem with other options, as of now, is that much of the language that Therian identities use is also being co-opted by the AI industry. "authorship" can be achieved by AI. LLMs are "creations". Indeed, much of the drivel coming from "AI ethicists" poses AI behaviour as indicative of "person"-hood, and of sapience.

I suspect that the standard will not move away from human.json, not out of exclusivity to Therian folks, but because of intentional (arguably malicious) characterisation of AI as people. Though, if I were to propose such a name change, perhaps one could focus on corporeality rather than identity. flesh.json? meat.json?

FWIW, one can, as of right now, specify the human.json to any name of their choosing. If one does not wish to associate themselves accordingly, I think it makes perfect sense to use, e.g.: `<link rel="human-json" href="/being.json">`. The problem with other options, as of now, is that much of the language that Therian identities use is also being co-opted by the AI industry. "authorship" can be achieved by AI. LLMs are "creations". Indeed, much of the drivel coming from "AI ethicists" poses AI behaviour as indicative of "person"-hood, and of sapience. I suspect that the standard will not move away from `human.json`, not out of exclusivity to Therian folks, but because of intentional (arguably malicious) characterisation of AI as people. Though, if I were to propose such a name change, perhaps one could focus on corporeality rather than identity. `flesh.json`? `meat.json`?

robida commented 2026年04月05日 00:23:24 +02:00

Thanks for the info, @mosssystem! I did not take this in consideration when I started the project and, TBH, this is the first time I've heard the term "therian". I chose the term "human" as a juxtaposition to "machine", and less as an identity.

As @addisoncrump mentioned, thanks to the discovery mechanism you can name the file whatever you want, but you still need to indicate it using rel="human-json", which is not ideal. I wouldn't be opposed to changing this in the future to person-json or something similar.

(Obligatory They're Made out of Meat.)

Thanks for the info, @mosssystem! I did not take this in consideration when I started the project and, TBH, this is the first time I've heard the term "therian". I chose the term "human" as a juxtaposition to "machine", and less as an identity. As @addisoncrump mentioned, thanks to the discovery mechanism you can name the file whatever you want, but you still need to indicate it using `rel="human-json"`, which is not ideal. I wouldn't be opposed to changing this in the future to `person-json` or something similar. (Obligatory [They're Made out of Meat](https://web.mit.edu/people/dpolicar/writing/prose/text/thinkingMeat.html).)

mosssystem commented 2026年04月06日 08:41:22 +02:00

All reasonable points.
@addisoncrump
I agree that the cooption of the terms for authorship and creation is sucky, and likely most easy picks in that category have already got some of that muck on them already. Deliberately muddied waters.
(Not every Therian identifies as a living creature, so as much as I am laughing at meat.json, I think there's a few issues with it.)

The filename is a good point, and might be something actionable to include in documentation.

@robida No worries, I didn't suspect you chose the name out of any malice or anything. It's a solid one, a punchy one, and I don't know if there's a better one. I just wanted to make sure that others not usually represented got consideration before ruling that out.

All reasonable points. @addisoncrump I agree that the cooption of the terms for authorship and creation is sucky, and likely most easy picks in that category have already got some of that muck on them already. Deliberately muddied waters. (Not every Therian identifies as a living creature, so as much as I am laughing at meat.json, I think there's a few issues with it.) The filename is a good point, and might be something actionable to include in documentation. @robida No worries, I didn't suspect you chose the name out of any malice or anything. It's a solid one, a punchy one, and I don't know if there's a better one. I just wanted to make sure that others not usually represented got consideration before ruling that out.

fluffy-critter commented 2026年04月08日 19:40:59 +02:00

Thinking of therian and other alterhuman identities is why I specifically refer to my anti-AI measures as a "sentience test" rather than a "human test," although I also know some folks who do not consider themselves sentient, and I'm not sure how to respond to that one.

I don't identify as therian or nonhuman myself but I do bristle a bit on behalf of my therian friends whenever I see the term "human" applied to meaning "non-automated bot" or the like.

Thinking of therian and other alterhuman identities is why I specifically refer to my anti-AI measures as a "sentience test" rather than a "human test," although I also know some folks who do not consider themselves sentient, and I'm not sure how to respond to that one. I don't identify as therian or nonhuman myself but I do bristle a bit on behalf of my therian friends whenever I see the term "human" applied to meaning "non-automated bot" or the like.

addisoncrump commented 2026年04月14日 00:29:52 +02:00

as much as I am laughing at meat.json, I think there's a few issues with it

I just learned about the AI hardware which includes rat brain cells. So corporeality is also out...

> as much as I am laughing at meat.json, I think there's a few issues with it I just learned about the AI hardware which includes rat brain cells. So corporeality is also out...

Wolkensteine commented 2026年04月23日 19:22:13 +02:00

Since human.json also encourages people to have an /ai page to declare their stance on the matter, I would suggest changing the name maybe to something that doesn't directly reflect on the individual who created the content, as it would probably be difficult to incorporate everyone (for example I have also no idea how to incorporate people who don't want to be seen sentient as @fluffy-critter mentioned), but to name it in a way it shows what it is not. The first thing that came to mind for me was noSlop.json, so you can indicate high-quality content. But I know that slop is very vague and thereby probably not the best word for this.

Since human.json also encourages people to have an /ai page to declare their stance on the matter, I would suggest changing the name maybe to something that doesn't directly reflect on the individual who created the content, as it would probably be difficult to incorporate everyone (for example I have also no idea how to incorporate people who don't want to be seen sentient as @fluffy-critter mentioned), but to name it in a way it shows what it is not. The first thing that came to mind for me was noSlop.json, so you can indicate high-quality content. But I know that slop is very vague and thereby probably not the best word for this.

fluffy-critter commented 2026年04月23日 19:29:29 +02:00

I think in the bigger picture, the aspect this idea is trying to get at is establishing a non-algorithmic source of authorship, right? Maybe the nomenclature can be based on that.

I think in the bigger picture, the aspect this idea is trying to get at is establishing a non-algorithmic source of authorship, right? Maybe the nomenclature can be based on that.

bannmann commented 2026年05月14日 21:21:54 +02:00

Hmm, may we assume that at least most people who don't identify as human do identify as "biological" or "organic"? If so, those may be appropriate terms that hopefully will not be co-opted by the "AI" industry any time soon.

Hmm, may we assume that at least _most_ people who don't identify as human _do_ identify as "biological" or "organic"? If so, those may be appropriate terms that hopefully will not be co-opted by the "AI" industry any time soon.

fluffy-critter commented 2026年05月14日 21:35:24 +02:00

I know a lot of folks who identify as robots or automata.

Identity is weird.

I know a lot of folks who identify as robots or automata. Identity is weird.

joepie91 commented 2026年05月24日 16:11:55 +02:00

This will probably not be punchy enough, but I'd be thinking in the direction of self-authored.json. Focusing not just on the 'authored' part (as that gets co-opted very easily), but on the relationship of who authored it, without making any further assertions about the identity of said 'who', just that it is the same entity as the operator of the site. Perhaps someone else can come up with a punchier equivalent of the same concept?

The main flaw I can see with this approach is AI bros claiming that this should cover slop machines generating entire sites including the self-authored.json. But at some point there's also the question of how 'watertight' the name really needs to be when there's documentation that makes clear what the 'spirit of the law' is, so to say. Especially taking into account that it's generally impossible to perfectly defeat co-opting through wording choice alone, as centuries of activist history show.

So perhaps all that's needed is a name that would make co-opting attempts look suspect (so that people look at the definition), rather than defeating them out of the gate with a single word. I think that'd also make it easier to find an inclusive name that works.

This will probably not be punchy enough, but I'd be thinking in the direction of `self-authored.json`. Focusing not just on the 'authored' part (as that gets co-opted very easily), but on the relationship of *who* authored it, without making any further assertions about the identity of said 'who', just that it is the same entity as the operator of the site. Perhaps someone else can come up with a punchier equivalent of the same concept? The main flaw I can see with this approach is AI bros claiming that this should cover slop machines generating entire sites including the `self-authored.json`. But at some point there's also the question of how 'watertight' the name really needs to be when there's documentation that makes clear what the 'spirit of the law' is, so to say. Especially taking into account that it's generally impossible to perfectly defeat co-opting through wording choice alone, as centuries of activist history show. So perhaps all that's needed is a name that would make co-opting attempts look *suspect* (so that people look at the definition), rather than defeating them out of the gate with a single word. I think that'd also make it easier to find an inclusive name that works.