Confused about silicon laws and identity theft

[Raven776]

So once a while back, someone's job was to steal a functional AI when there was none on board. They moidered me, brained me, and stuck me in an AI unit in maint. I turned off the APC as quickly as I could to try and stop them from uploading laws, but I was too slow.

Here's the only problem, they also stole my identity and lawed me that I (previously human I) was the only human and to listen to them.

So obviously this brain in this AI was the brain of Captain McCaptain, and the CMO was posing as Captain McCaptain, so how do you handle that? Since they have the face and genes of Captain McCaptain, does that make them the new one and therefore the only human?

Besides showing some initial obvious confusion over who the real Captain is, hinting that I MIGHT kill them just so they'd change my laws into something more reasonable, I just let it all go.

So here's my question, what counts as identity for an AI? When you have a one hundred percent certainty that someone isn't who they say they are or look like, can you as an AI moider them if you're keyed to only protect one person?

[Lo6a4evskiy]

If the law was "Captain McCaptain is the only human", then every Captain McCaptain is human and nobody else is. It doesn't matter if they're changeling or not. Law explicitly states that they are human. You must treat them as such.

Think ten times before dicking someone over just because they're an antagonist. AIs are not validhunters.

[peoplearestrange]

Essentially when someone one humans the AI with a name, anyone with that name can be counted. This can work well as an Anti-counter (though its basicly never used, and its not that simple to work out who the actual one human is).

You take who they are Identity wise from examining or hovering you cursor over them. e.g:

1. Sam Smith is the only human.

This is Sam Smith - The AI should follow their requests.

This is David McDaveson (As Sam Smith) - Their true Identity has been revealed to the player (the AI) due to taking their mask of when wearing someone else's ID etc etc and therefore are NOT actually Sam Smith as far as the laws go.


At least this is my interpretation.

[Kelenius]

But in this particular case, the AI knows the true identity of the masked person.

[Steelpoint]

Which is a fair point, under normal circumstances the AI would have to obey its law. However this AI in particular knows for 100% certainty that Captain McCaptain is not actually Captain McCaptain.

In my opinion, this would fall under the same circumstance of the AI knowing someone is a Changeling. Sure the ling might be the HoP, but its still a ling.

Since we don't enforce memory wipes on characters being turned into a AI or Cyborg, there is no reason for this new AI to know what's up.

[Pandarsenic]

Now here's the complicated part - the guy apparently stole the UI and UE of Captain McCaptain at genetics to BECOME a Captain McCaptain. Are they still who they were before, or are they Captain McCaptain now?

Will post more and better later once people have put in some input on the above, but I'm inclined to say if they go that far the identity and name are theirs.

[Steelpoint]

Its a conflict, on the one hand the AI knows this Captain is not the actual Captain.

On the other hand, this fake Captain is actually the Captain down to the DNA level.

In my opinion, while the AI does know the Captain is fake, he is technically the Captain. Thus is afforded the protection of a human under the lawset. Which means the AI cannot reveal the true identity of the fake Captain since that would cause human harm.

[Scott]

Now here's the complicated part - the guy apparently stole the UI and UE of Captain McCaptain at genetics to BECOME a Captain McCaptain. Are they still who they were before, or are they Captain McCaptain now?

Will post more and better later once people have put in some input on the above, but I'm inclined to say if they go that far the identity and name are theirs.

I think the AI should identify people by their genetic makeup (when it can see the humans faces), like when a changeling is exposed it still looks like a human, but the AI knows that, genetically, that shit ain't human. So when a person alters their DNA to become a different person, the AI should consider that modified human to be different, now that it has a new genetic makeup, when following One Human laws.

What is in doubt, to me, is, when pure Asimov, whether or not the AI should consider that genetically different person to be the same person as before in case of that person being a killer. Is the AI going to act dumb and let you go free if you become a different person or can it still consider you a threat?

[Alex Crimson]

Do characters retain their memories when they become Silicon? I thought not. At the very least, i didnt think they were allowed to act on them. If a Traitor Robo kills someone and borgs them, can the borg choose a Sec module and instantly arrest the Traitor?

In this situation, if i were the AI, i wouldve gone along with the fake captain until he was revealed to be a fake, ignoring my IC knowledge and treating it as a new round as soon as i became the AI.

If he goes as far as to change his genetics, then i think he has earned the right to become the "real" captain.

[Scott]

Do characters retain their memories when they become Silicon? I thought not. At the very least, i didnt think they were allowed to act on them. If a Traitor Robo kills someone and borgs them, can the borg choose a Sec module and instantly arrest the Traitor?

In this situation, if i were the AI, i wouldve gone along with the fake captain until he was revealed to be a fake, ignoring my IC knowledge and treating it as a new round as soon as i became the AI.

If he goes as far as to change his genetics, then i think he has earned the right to become the "real" captain.

It is the same brain, so yes, memories are retained. At least I have always retained them and never got into trouble for it. You only cannot bring info from the afterlife into your role play, as you can see dead chat when you're a popped brain.

[Lo6a4evskiy]

But in this particular case, the AI knows the true identity of the masked person.

What's "true identity"? The identity with which you start the round? Why? Why should AI even care about that?

This guy has an appearance of Captain McCaptain and a voice of Captain McCaptain, even exact fucking DNA of Captain McCaptain, I see no reason for AI not to obey.

Don't forget that lawset comes before everything else IC, even if there's a law that states "Changeling McChangeling is human" and AI sees that Changeling McChangeling has armblade and ling armor and is using ling abilities all the time, Changeling McChangeling is still human.

[peoplearestrange]

You only cannot bring info from the afterlife into your role play, as you can see dead chat when you're a popped brain.

This brings up something that I've wondered for awhile now.
WHY do brains get to see dead chat? I mean are we supposed to assume the operation kills the brain, then the MMI magically restores sentience/brain activity? Or do we assume the operation somehow keeps the brain alive, just without communication or "input" from the world around until the MMI is used on it?

[Scott]

It's because brains are actually omniscient.

[Raven776]

Two notes for this particular scenario.

First speaking for myself when I was playing through it: I let them get away with it and aided them in their escape, partially because it was a clever thing to do and partially because I was unsure of the rules.

Second note: The law was "X is the only human." If we start to define it genetically, then a traitor who steals another face through genetics and informs the AI can them be screwed over. If we define it by round start, then technically I as the AI was the only human since I was the true X. I knew with 100% certainty that they were not X.


And another bit, borgs retain their memories as far as I can tell. Borgs can be made to go through interrogations, can retain personality traits, but the reason that a traitor borged cyborg can't go sec and immediately arrest someone before being subverted is because that could lead to human harm. Unless they KNOW they're going to be subverted in which case arrest away because it's your JOB as a split second, probably flashed or EMPed secborg to do that.

Edit: I'd actually kinda like to see it ruled that anyone with the name counts as the person. It would be so much fun to see a thousand gas masked people ordering a subverted AI around through PDAs with the traitor's name on it.

[Cipher3]

Second note: The law was "X is the only human." If we start to define it genetically, then a traitor who steals another face through genetics and informs the AI can them be screwed over. If we define it by round start, then technically I as the AI was the only human since I was the true X. I knew with 100% certainty that they were not X.

The problem here is that you're thinking of their identity as absolute. They WEREN'T X, but it could well be that they BECAME X. That's not determined policy, since I've never seen this scenario before, but it is very much a reasonable case to make that they became Captain McCaptain when they took on everything about him.

[Raven776]

Very true, but there are no rules on what constitutes identity as far as AIs are concerned.

Here's a question: What if a traitor's identity was found out, it was obvious he was onehumaned to the AI, and instead of taking the few simple steps it would take to fix the AI, the crew goes through a needlessly convoluted set of steps to steal his identity and mass pose as him? If the original Traitor was still known absolutely to the AI (he was in the AI's core the entire time), then would all of the new traitors be human in the AI's eyes too?

I'd say yes because I find it hilarious, but everyone's sense of humor isn't mine.

[Deuryn]

I might be mistaken, but I think we do have some previous things to go on with this. Back in the day of retrovirus, or whatever it was that turned everyone into the same person, we had a one-human that was everyone.

The law is defined as "Only Player X is human" since everyone is Player X, everyone is human. We're defining by name, so I think, and we'll wait for Hamburger Helper to agree or disagree before I declare anything for certain but, here's the tentative ruling.

If you can successfully steal someone's identity on a genetic scale, you can be the one human as that person, even if the AI would know you're really not them. You are that person for all intents and purposes within the game.

This brings up something that I've wondered for awhile now.
WHY do brains get to see dead chat? I mean are we supposed to assume the operation kills the brain, then the MMI magically restores sentience/brain activity? Or do we assume the operation somehow keeps the brain alive, just without communication or "input" from the world around until the MMI is used on it?

I think it's from the code. If I remember right, brains and MMIs are objects, not mobs. Because you aren't in a deadchat disabled mob, you get deadchat. I think.

[Scott]

If you can successfully steal someone's identity on a genetic scale, you can be the one human as that person, even if the AI would know you're really not them. You are that person for all intents and purposes within the game.

What about a regular Asimov AI, is it supposed to stop considering a known killer as a threat when the killer alters his DNA to become a "new" person?

[Lo6a4evskiy]

What about a regular Asimov AI, is it supposed to stop considering a known killer as a threat when the killer alters his DNA to become a "new" person?

Why would it?

[Pandarsenic]

Short answer is if "(Only) X is/are human" is a law, any and all X are human. However, you can treat individual instances of that X differently based on their actions.

Long answer is I will go on a huge type/token rant to explain this if people need.

[Raven776]

I kinda wanna hear the rant.

[callanrockslol]

It doesn't matter who you were, at all, you really only have to check the person has the right name and that they aren't a guy with a mask and ID.

Gameplay perspective, if they have become you and their name matches up then for the sake of one human they are you even if they were someone else before, if its just a guy in a mask then they are until you can prove otherwise. If they are a ling THEIR STILL HUMAN FORGET WHAT I WROTE BEFORE.

IC perspective, a deep fundamental question about the nature of humanity and individual identity that can be completely ignored because they are the person that your law says is human.

It can go a lot deeper but since AI's are expected to literally follow their laws it helps make it work.

[Cipher3]

If they are a ling ignore both conditions and remove like kebab.

*If it is x, fulfilling all conditions, and X is the one human, then x pulling an armblade does not make him nonhuman. X turning into Y makes him Y, while X is still the only human, therefore Y is not human.

[Kelenius]

But

The law says that only X /is/ human.

Is, not are. As in there is only one of them.

[Kangaraptor]

If the law specifically states x is human, x is human even if they appear nonhuman because the definition of 'human' is attributed directly to a name and face, not their extensive genetic makeup (as is inferred by the stock definition of 'human' by the AI).

Now, if x goes ahead and changes shape into y, I think it's safe for the AI to say that y is no longer human because... they're not x. Simply put: the definition of humanity is in their appearance and name, once they no longer have that appearance and name they're no longer human.

[callanrockslol]

If they are a ling ignore both conditions and remove like kebab.

*If it is x, fulfilling all conditions, and X is the one human, then x pulling an armblade does not make him nonhuman. X turning into Y makes him Y, while X is still the only human, therefore Y is not human.

Whoops just realized what I wrote.

YES EVEN IF THEY ARE A LING THEY ARE A HUMAN UNDER ONE HUMAN UNTIL THEY CHANGE NAMES FORGET I SAID THE PACHYDERM

But

The law says that only X /is/ human.

Is, not are. As in there is only one of them.

They would just be the same person in multiple places at once obviously, as long as the name fits.

If the law specifically states x is human, x is human even if they appear nonhuman because the definition of 'human' is attributed directly to a name and face, not their extensive genetic makeup (as is inferred by the stock definition of 'human' by the AI).

Now, if x goes ahead and changes shape into y, I think it's safe for the AI to say that y is no longer human because... they're not x. Simply put: the definition of humanity is in their appearance and name, once they no longer have that appearance and name they're no longer human.

I think its just their name actually, doesn't really matter what they look like if their name matches up, people get fucked up genetic damage and surgery and such and are still the same person.

[Kangaraptor]

If the law specifically states x is human, x is human even if they appear nonhuman because the definition of 'human' is attributed directly to a name and face, not their extensive genetic makeup (as is inferred by the stock definition of 'human' by the AI)
Now, if x goes ahead and changes shape into y, I think it's safe for the AI to say that y is no longer human because... they're not x. Simply put: the definition of humanity is in their appearance and name, once they no longer have that appearance and name they're no longer human.

I think its just their name actually, doesn't really matter what they look like if their name matches up, people get fucked up genetic damage and surgery and such and are still the same person.

it wouldn't be, because if someone is unmasked and is wearing someone else's ID it'll tell you (even as the AI). So it'd say Y mcYerson (as x) (which means Y isn't X), whereas if they were masked or otherwise concealing their true identity then sure you'd go by voice alone.

[callanrockslol]

Fucking uselees android phone, can't finish the sentance


If it said they are Fagit Griffin (as valiant mccaptain) they would be fagit griffin fr law purposes, if they had a mask and showed up asj valiant mccaptain, then when they talked came uu

[cedarbridge]

You only cannot bring info from the afterlife into your role play, as you can see dead chat when you're a popped brain.

This brings up something that I've wondered for awhile now.
WHY do brains get to see dead chat? I mean are we supposed to assume the operation kills the brain, then the MMI magically restores sentience/brain activity? Or do we assume the operation somehow keeps the brain alive, just without communication or "input" from the world around until the MMI is used on it?

Consider that AIs traditionally recognize borging as a harmful act. This is because the human dies when their brain is removed. I'd imagine this is an extension of the above.

[callanrockslol]

There isn't anything harmful about dying, it's just that it tends to be done in harmful ways, borging used to do no damage and wasnt considered harmful at all, but now it does like 80 brute on the second last step when you cut they skull open.

Cessation of Life=/=Harm. It just tends to happen in harmful ways.

[peoplearestrange]

Consider that AIs traditionally recognize borging as a harmful act. This is because the human dies when their brain is removed. I'd imagine this is an extension of the above.

The human may die as a whole, yes, but the brain does not, therefore memories are probably not forgotten either. This is the way I have always thought of it. Think of it in a Robocop way, you were human once, memories are there, vague sometimes, overridden by its lawset, but still there. Least that's the way I've always RPed it.

Therefore I would say that the AI would probably remember who it was, and seeing someone as it self, unless it saw the process of identity stealing, would simply confuse the shit out of it. I wouldn't of thought it could jump to the conclusion that it is stolen identity, but certainly worth containing and questioning over this "glitch".

Thing is its iffy grounds for a circumstance that doesn't really occur that often. Due to the simple set of laws there will always be exceptions, loopholes and missing info on what do to when. This is the flaw (and why its interesting) with asimov's lawset.

[Pandarsenic]

I suppose it'd be too much to ask "What's most fun for everyone"?

[Aurx]

I suppose it'd be too much to ask "What's most fun for everyone"?

I'd love to have a polling method to get accurate results from everyone.
I'd love to have a way to quantify fun.

Care to share your methods?

[Pandarsenic]

I don't mean in general, I mean on a case-by-case basis. If you can defend how you made the round more fun for other people with your choice in context, I'd probably just allow it.