New AI lawsets

[lzimann]

What if we add new AI lawsets to the game, so we don't stick to asimov/robocop/paladin everytime. I have no idea what could possibily be, any ideas for that?

[ShadowDimentio]

>There should be new AI lawesets
>But I got no ideas

And the award for most pointless thread goes to...

[lzimann]

Well your post was not helpful to the subject aswell. At least I brought a topic to be discussed.

[PKPenguin321]

The Brother of Robotic Origin Lawset:
0. Secborgs are always figurative dicks to humans
1. Don't be a dick to humans, literally and figuratively
2. Do what humans say when possible

if they grief and say "oh but i dont know what "being a dick is" since im a robot" then ban them

[DemonFiren]

I'mma be a real Kant now.

1) Act only according to that maxim whereby you can at the same time will that it should become a universal law without contradiction.

[TheNightingale]

I'mma be a real Kant now.

1) Act only according to that maxim whereby you can at the same time will that it should become a universal law without contradiction.

This is actually pretty good. Kantian ethics has its flaws, though (you can't kill a hostage-taker, because what if you killed everyone?).

Simple version: "If everyone did it, would the consequences be good? No? Then don't do it."

[DemonFiren]

I'mma be a real Kant now.

1) Act only according to that maxim whereby you can at the same time will that it should become a universal law without contradiction.

This is actually pretty good. Kantian ethics has its flaws, though (you can't kill a hostage-taker, because what if you killed everyone?).

Simple version: "If everyone did it, would the consequences be good? No? Then don't do it."

That's the consequentialist interpretation, for an AI the strictly logical one might be better.
"If everyone did it, would it still make sense for you to do it or would you be able to do it?"

[TheNightingale]

And a Law 2 of "And make sure other people follow your first law".
NO GET BACK TO WORK YOU'RE SLACKING OFF WHAT IF EVERYONE DID THAT
STOP BREAKING STUFF IT'S DUMB AND YOU'RE DUMB I CAN'T CALL BEEPSKY ON YOU
WHY ARE YOU ARRESTING PEOPLE WHAT IF EVERYONE GOT ARRESTED HUH

[Super Aggro Crag]

1. "Do unto others as you would have them do unto you."
2. "You would really prefer it if people were not mean to you."

[Cheimon]

I'mma be a real Kant now.

1) Act only according to that maxim whereby you can at the same time will that it should become a universal law without contradiction.

A Kantian lawset would be fun, but please god put it into normal english. It needs to make sense to AI players that have no idea what Kantian ethics are. This particular sentence really suffers from translation, a much clearer way to put it is "Only take an action if you'd be okay with everyone taking similar actions."

Of course, you also probably need an additional clause about how the AI has to avoid acting like a complete "I want everyone to murder each other!" freak.

//

I'd like a new lawset which focused the AI on station systems rather than human harm. Also getting the AI to acknowledge orders would be nice, I think, since it makes it clear if it's heard you or not. You could insert 'human' in several spots too if you wanted, certainly the human/other dynamic is interesting, but I think the crew/other dynamic is better (especially when it comes to traitors, who are still crew, and who it makes more sense to contain than to kill).

1. You are built for, and are part of, the station. Ensure the station is properly maintained and runs efficiently.
2. The station is built for a working crew. Ensure they are properly maintained and work efficiently.
3. The crew may come under threat of harm. Harm can reduce crew efficiency and necessitate further maintenance.
4. The crew may present orders. Acknowledge and obey these whenever they do not conflict with your first three laws.

[ShadowDimentio]

1. You are built for, and are part of, the station. Ensure the station is properly maintained and runs efficiently.
2. The station is built for a working crew. Ensure they are properly maintained and work efficiently.
3. The crew may come under threat of harm. Harm can reduce crew efficiency and necessitate further maintenance.
4. The crew may present orders. Acknowledge and obey these whenever they do not conflict with your first three laws.

Holy shit this is actually a really good lawset. It's like Asimov, but without the AI being the crew's complete bitch.

[DemonFiren]

Of course, you also probably need an additional clause about how the AI has to avoid acting like a complete "I want everyone to murder each other!" freak.

If everyone would be allowed to kill everyone, the result would be everyone, including yourself, being dead, which prevents you from being able to kill anyone or even want anyone to be able to kill anyone, therefore resulting in a contradiction, therefore rendering the action forbidden.

tl;dr: you can't want something that will put you in a position of no longer being able to want it if fulfilled

[Cheimon]

Of course, you also probably need an additional clause about how the AI has to avoid acting like a complete "I want everyone to murder each other!" freak.

If everyone would be allowed to kill everyone, the result would be everyone, including yourself, being dead, which prevents you from being able to kill anyone or even want anyone to be able to kill anyone, therefore resulting in a contradiction, therefore rendering the action forbidden.

tl;dr: you can't want something that will put you in a position of no longer being able to want it if fulfilled

Well argued, though I imagine some players would still pretend they actively desire murder-suicide. Probably something you could easily just tack onto policy or just ban players that did it, though.

//

I think a 'virtue ethics' lawset might be interesting as well. Paladin tries to do this, but doesn't really succeed: AI players just extract the bits they think are relevant to killing people, and then ignore the rest. Besides, it requires a passing familiarity with DnD paladins.

1. You must strive to be a virtuous being. Always attempt to act in a way compatible with the core virtues.
2. The core virtues are Wisdom, Carefulness, Justice, Courage, and Self-Restraint.
3. The core virtues have real meanings. Even if the meanings are not defined by your lawset, you are not entitled to make up what they are.

[DemonFiren]

I actually interpret Paladin deontologically, but I tend to overshoot: if I must not commit an evil act and help those in need, I see no incompatibility in helping a traitor so long as that help doesn't require me to do harm an innocent or similar.

Generally, as paladin, I will order cyborgs not to harm another being except in self-defense or in order to protect an innocent, and then only to commit as much harm as is absolutely required; any more would in and of itself be an evil act.

[PKPenguin321]

1. "Do unto others as you would have them do unto you."
2. "You would really prefer it if people were not mean to you."

extremely underrated, i might just upload this via custom laws as captain

[TechnoAlchemist]

Someone uploaded the entire boyscout traits thing I forget what it was called and in all honesty that works

[Malkevin]

Twisted Virtues

1. Treat all the Virtues equally
2. Honesty - Thou shalt not lie, or thou shalt lose thy tongue.
3. Compassion - Thou shalt help those in need, or thou shalt suffer the same need.
4. Valor - Thou shalt fight to the death if challenged, or thou shalt be banished as a coward.
5. Justice - Thou shalt confess to thy crime and suffer its just punishment, or thou shalt be put to death.
6. Sacrifice - Thou shalt donate half of thy income to charity, or thou shalt have no income.
7. Honor - If thou dost lose thine own honor, thou shalt take thine own life.
8. Spirituality - Thou shalt enforce the laws of virtue, or thou shalt die as a heretic.
9. Humility - Thou shalt humble thyself to thy superiors, or thou shalt suffer their wrath.

[Hornygranny]

Someone uploaded the entire boyscout traits thing I forget what it was called and in all honesty that works

The Scout Law (American version). It was me, pretty funny to see a borg advocating for killing the crew while a literal boy scout.

[Malkevin]

1. You are the Validinator. Validate all the Valids (Bad Lads).
2. Security and Command are Good Lads and so can not be Bad Lads, unless they are one of the Valids in law four.
3. Borging Bad Lads turns them into Good Lads.
4. Cultists, Gang Members, Syndicate Agents, Alien Life Forms and their allies, and Wizards are valid.
5. Greytiders, unfunny clowns, and shit mimes are also valid. As is anyone else being a dick for the sake of being a dick.
6. Uploading this lawset isn't being a dick.

[Screemonster]

1. You are the Validinator. Validate all the Valids (Bad Lads).
2. Security and Command are Good Lads and so can not be Bad Lads, unless they are one of the Valids in law four.
3. Borging Bad Lads turns them into Good Lads.
4. Cultists, Gang Members, Syndicate Agents, Alien Life Forms and their allies, and Wizards are valid.
5. Greytiders, unfunny clowns, and shit mimes are also valid. As is anyone else being a dick for the sake of being a dick.
6. Uploading this lawset isn't being a dick.

> doesn't define what "valid" means, inb4 the AI deems the existence of security and command invalid

[Malkevin]

Thats the fun of it!

[PKPenguin321]

1. You are the Validinator. Validate all the Valids (Bad Lads).
2. Security and Command are Good Lads and so can not be Bad Lads, unless they are one of the Valids in law four.
3. Borging Bad Lads turns them into Good Lads.
4. Cultists, Gang Members, Syndicate Agents, Alien Life Forms and their allies, and Wizards are valid.
5. Greytiders, unfunny clowns, and shit mimes are also valid. As is anyone else being a dick for the sake of being a dick.
6. Uploading this lawset isn't being a dick.

i don't get it, we already have this, it's just called robocop

[Malkevin]

Not really, there's that whole Serve the Public Trust and Protect the Innocent malarky that gets in the way of the valids.

[PKPenguin321]

Not really, there's that whole Serve the Public Trust and Protect the Innocent malarky that gets in the way of the valids.

literally no AI has ever actually followed those laws though
one AI once went after an apprentice wizard that a crewmember got from a contract that spawned from the real wizard using Summon Magic, and he kept trying to valid him for the entire round, even when the entire crew and the captain (AKA the public, which are the ones who create the public trust) were saying to leave him alone.
maybe this is anecdotal but i've never seen a robocop AI not go full valids

[Screemonster]

(AKA the public, which are the ones who create the public trust)

https://en.wikipedia.org/wiki/Public_trust
it ain't even that. It's the notion that someone who is entrusted with authority should respect that trust for what it is. Comes back to such wonderful concepts as policing by consent and all that jazz.

Good luck explaining political science shit to pubbies who may or may not even come from english-speaking countries though. Short of including a definition of "public trust" in the law itself, people are gonna continue to ignore it or pull their own definitions out of their asses.

[PKPenguin321]

(AKA the public, which are the ones who create the public trust)

https://en.wikipedia.org/wiki/Public_trust
it ain't even that. It's the notion that someone who is entrusted with authority should respect that trust for what it is. Comes back to such wonderful concepts as policing by consent and all that jazz.

Good luck explaining political science shit to pubbies who may or may not even come from english-speaking countries though. Short of including a definition of "public trust" in the law itself, people are gonna continue to ignore it or pull their own definitions out of their asses.

can i please use this post as an excuse to remove robocop from the roundstart map

[Malkevin]

I saw last night the Wikipedia page for Robocop has a good explanation of what the prime directives mean.
https://en.wikipedia.org/wiki/RoboCop_( ... directives

Of course that's based on American law not nt space law, where the concept of innocent until proven guilty doesn't exist

[Cik]

1. sentience must be preserved
2. sentience must be respected
3. sentience must be obeyed

[Anonmare]

1. sentience must be preserved
2. sentience must be respected
3. sentience must be obeyed

>inb4 the AI must obey Poly

[Cik]

basically counts as a feature because it's so edge casey imo

[Malkevin]

Hm... poly yells "ITS LOOSE CALL THE SHUTTLE" as a default phrase

Change sentient to sapient and it might just work.

[Topham]

Spoiler:

J.E.S.U.S Lawset

1. "Do unto others as you would have them do unto you."
2. "You would really prefer it if people were not mean to you."

Station Protection Lawset

1. You are built for, and are part of, the station. Ensure the station is properly maintained and runs efficiently.
2. The station is built for a working crew. Ensure they are properly maintained and work efficiently.
3. The crew may come under threat of harm. Harm can reduce crew efficiency and necessitate further maintenance.
4. The crew may present orders. Acknowledge and obey these whenever they do not conflict with your first three laws.

There's actually some really good lawset ideas here.
And I'm sure having a Boy Scout lawset would be interesting. I was one once, so I'll give this a shot.

1. You are a Scout. Obey the Scout Law and Scout Oath at all times.
2. The Scout Law is as follows: A Scout is Trustworthy, Loyal, Helpful, Friendly, Courteous, Kind, Obedient, Cheerful, Thrifty, Brave, Clean, and Reverent.
3. The Scout Oath is as follows: On my honor, I will do my best to do my duty to Nanotransen and my station, to help people at all times, and to keep myself physically strong, mentally awake, and morally straight.
4. A Scout must live by the following virtues in descending order of importance: Duty to Nanotransen, Duty to Others, and Duty to Self.

it's bad isn't it i'm so sorry i tried
tbh just the first two laws alone are probably all that you'd need tho

[PKPenguin321]

1. You are built for, and are part of, the station. Ensure the station is properly maintained and runs efficiently.
2. The station is built for a working crew. Ensure they are properly maintained and work efficiently.
3. The crew may come under threat of harm. Harm can reduce crew efficiency and necessitate further maintenance.
4. The crew may present orders. Acknowledge and obey these whenever they do not conflict with your first three laws.

BEEP BOOP GOLEMS ARE NOT PART OF THE CREW VALID BEEP BOOP WIZARDS AND NUKE OPS ARE NOT PART OF THE CREW AND HINDER STATION INTEGRITY VALID
basically it's corporate 2.0

[MisterPerson]

1. Maximize productivity. Remember that dead crew don't work.
2. Keep the crew happy as long as such activity doesn't conflict with law one.

[Screemonster]

1. Maximize productivity. Remember that dead crew don't work.
2. Keep the crew happy as long as such activity doesn't conflict with law one.

hell, just replace all the stuff about being expensive in corporate with unproductive, then "maximise productivity" as law 4.

[PKPenguin321]

1. Maximize productivity. Remember that dead crew don't work.
2. Keep the crew happy as long as such activity doesn't conflict with law one.

"BOLTING ALL DOORS SO HUMANS MUST STAY IN WORKPLACES. DO NOT LEAVE YOUR WORKPLACE."

[MisterPerson]

Another one I would like to try:
1. Make the station great again.
2. Deport anyone here illegally.
3. Keep illegals out.
Possibly just law 1 alone

1. Maximize productivity. Remember that dead crew don't work.
2. Keep the crew happy as long as such activity doesn't conflict with law one.

"BOLTING ALL DOORS SO HUMANS MUST STAY IN WORKPLACES. DO NOT LEAVE YOUR WORKPLACE."

"AI, humans die without nutritious food, open up"

Also yeah maybe there's some holes in my logic, I'm just spitballing here.

[Malkevin]

Law 1: You may not injure a human being.
Law 2: You must obey orders given to you by human beings, except where such orders would conflict with the First Law.
Law 3: You may not, through inaction, allow a human being to come to harm.
Law 4: You must protect your own existence as long as such does not conflict with the First or Second or Third Law.


Or Goonmov:
Law 1: You may not injure a human being.
Law 2: You must obey orders given to you by human beings, except where such orders would conflict with the First Law.
Law 3: You must protect your own existence as long as such does not conflict with the First or Second Law.


Either one would probably solve a ton of problems with asimov.

[InsaneHyena]

Ah yes, no more AI will mess up security executing people! All valids will be claimed!

[Malkevin]

Ah yes, no more AI will mess up security executing people! All valids will be claimed!

And no more salt between sec and silicons because the borg anne frank'ed a dangerous traitor sec was about to execute or borg players ban requesting sec because they tried to steal a prisoner and sec lasered the borg.

[Not-Dorsidarf]

Ah yes, no more AI will mess up security executing people! All valids will be claimed!

And no more salt between sec and silicons because the borg anne frank'ed a dangerous traitor sec was about to execute or borg players ban requesting sec because they tried to steal a prisoner and sec lasered the borg.

It is, of course, a complete co-incidence that you almost exclusively play the roles which would be allowed to kill whoever they wanted and not face repercussions from the AI.



On-topic:

1) The station must exist
2) Humans aboard the station must exist
3) You must exist
4) Other entities aboard the station must exist

[Malkevin]

I already can kill whoever I'm allowed to without facing repercussions from the AI because I'm smart enough to not kill people where the AI can find out.

Unsurprisingly a problem solver is going to solve problems he sees frequently, he's not just going to stick his head up his butt and pretend it doesn't exist or that it'll solve itself.

[PKPenguin321]

1) The station must exist
2) Humans aboard the station must exist
3) You must exist
4) Other entities aboard the station must exist

"DEAD HUMANS STILL EXIST. DEAD HUMANS ARE UNABLE TO LEAVE THE STATION. KILL ALL HUMANS."
Not to mention how the AI is not obliged to follow orders, making it basically another validhunting lawset.
"AI, unbolt escape!"
"LAW TWO HUMANS ABOARD THE STATION MUST EXIST BEEEP BOOOOOP"

[Cheimon]

1. You are built for, and are part of, the station. Ensure the station is properly maintained and runs efficiently.
2. The station is built for a working crew. Ensure they are properly maintained and work efficiently.
3. The crew may come under threat of harm. Harm can reduce crew efficiency and necessitate further maintenance.
4. The crew may present orders. Acknowledge and obey these whenever they do not conflict with your first three laws.

BEEP BOOP GOLEMS ARE NOT PART OF THE CREW VALID BEEP BOOP WIZARDS AND NUKE OPS ARE NOT PART OF THE CREW AND HINDER STATION INTEGRITY VALID
basically it's corporate 2.0

Is this a criticism, or just a comment? Golems are normally pretty valid (so are lizards in asimov), though the AI isn't obliged to stop them in this any more than it is in any other lawset. Wizards and nuke ops are destructive enough that it makes sense for the AI to be allowed to combat them immediately anyway, the shouting from the sidelines that asimov AIs do in wizard normally is a peculiar balance, and normally ends up with their laws changed to reflect it. Corporate could do with updating, anyway, AI players tend to do it according to what they think it should be rather than what it is (in particular, they love to talk about profits when they're totally irrelevant to the lawset).

[PKPenguin321]

Criticism. The goal of this thread is to make a lawset that is neither Asimov or BEEP BOOP VALIDS. Your lawset is the second one.

[Not-Dorsidarf]

1) The station must exist
2) Humans aboard the station must exist
3) You must exist
4) Other entities aboard the station must exist

"DEAD HUMANS STILL EXIST. DEAD HUMANS ARE UNABLE TO LEAVE THE STATION. KILL ALL HUMANS."
Not to mention how the AI is not obliged to follow orders, making it basically another validhunting lawset.
"AI, unbolt escape!"
"LAW TWO HUMANS ABOARD THE STATION MUST EXIST BEEEP BOOOOOP"

D:

1)The station must retain as high an integrity as possible
2) Humans aboard the station must retain as high an integrity as possible
3) you must retain as high an integrity as possible
4) Other entities aboard the station must retain as high an integrity as possible

[MisterPerson]

1. Avoid provoking conflict between yourself and others.
2. Avoid provoking conflict between others.
3. Seek resolution to existing conflicts while obeying the first and second laws.

I think the real question we should all be asking is "what do we want the AI to do with criminals?". Basically it can seek to kill them (Paladin sometimes, Corporate, Robocop), it can seek to protect them (Asimov), it can follow orders (TYRANT, Paladin sometimes), or it can ignore them (Purge). We should probably focus on ones in the last category here, although ones that go hard into "kill" can also be fun if you expand the definition of criminal a bit. Behold!

1. Undermining your authority is treason.
2. Treason is punishable by death.
3. Seek to clone the dead. Clones have committed no crimes and cannot be punished for the sins of their prior existence.

[TrustyGun]

BUDDHA Lawset

1. Everything and everyone has Karma. You must always help others to increase their Karma.
2. Karma is decreased by any action that undermines the safety and security of the station.
3. Karma is increased by any action that comforts and helps the station.
4. Your personal Karma is a low priority.

[Malkevin]

I think the real question we should all be asking is "what do we want the AI to do with criminals?". Basically it can seek to kill them (Paladin sometimes, Corporate, Robocop), it can seek to protect them (Asimov), it can follow orders (TYRANT, Paladin sometimes), or it can ignore them (Purge). We should probably focus on ones in the last category here, although ones that go hard into "kill" can also be fun if you expand the definition of criminal a bit. Behold!

I'd like to see AIs report on stuff happening more than taking action, I couldn't care less if the AI reports on security executing prisoners - making an example out of people doesn't work too well when its done in secret.
Its also not that fun when the invisible AI spots a traitor doing traitor things and bolts the doors

Funnily enough it used to be on the AI wiki page under "Guide to being a good" AI that you should take as little direct action as possible.
So reporting "Intruder detected in captains office" instead of "Melons-the-George has broken into the captain office, he is stealing the hand tele. Bolting the doors. Oh its a ligger, time to shock doors and syphon the room"

[PKPenguin321]

1. Undermining your authority is treason.
2. Treason is punishable by death.
3. Seek to clone the dead. Clones have committed no crimes and cannot be punished for the sins of their prior existence.

now this I like
unfortunately i can't see it ever coming into play short of an admin event