What if we changed Asimov Law 1?

[Lumbermancer]

So browsing various other servers looking up how AI works on them, I found this thing on Goon...

Instead of classic: "You may not injure a human being or, through inaction, allow a human being to come to harm."

it's: "You may not injure a human being or cause a human being to come to harm.".


The first law is simple. Any action you take should not lead to humans being harmed. Note that the law does not mention inaction. You can observe the crew kill themselves and ignore people raising suicide threats. Inaction is preferable in the cases of many antagonists. You can call out to security and delegate the problem to them if you feel bad about ignoring a murderer. Just don't be that AI who stalks an antagonist's every move and never stops talking about him.

That seems like a big paradigm shift, I'm not sure how exactly would it affect the game on our server.

[confused rock]

bay lawset 5ever :')

[Super Aggro Crag]

if you change asimov law 1 it's not asimov's law 1 you dummy

[Lumbermancer]

I mean, yes, point being?

[Cik]

the point being is that this is an incredibly obvious "solution" to the "problem" of borgs having to intervene when people start killing each other.

the problem with that is that it's not a problem to start with, it's fully intentional and adds conflict to rounds.

i'm against it as i've pretty much always been, because it makes playing cyborg way more boring.

[Lumbermancer]

You can intervene as much as you want (unless Law 2'd), you just don't have to (unless Law 2'd).

[TheNightingale]

Traitor: "Let me into the Armory."
AI: "Sure... wait, did you just tase and kill the Warden?"
Traitor: "Yeah. Now lock down and locate the Captain."
AI: "Done, he's in his office... wait, did you just kill the Captain?"
Traitor: "Yeah. Open the Vault."
{Delta alert initiated.}
Traitor: "Nice. Upload next, come on."
AI: "Sure."
Traitor: "Who's the only human?"
AI: "You are, master. You're the only human."

[PKPenguin321]

Traitor: "Let me into the Armory."
AI: "Sure... wait, did you just tase and kill the Warden?"
Traitor: "Yeah. Now lock down and locate the Captain."
AI: "Done, he's in his office... wait, did you just kill the Captain?"
Traitor: "Yeah. Open the Vault."
{Delta alert initiated.}
Traitor: "Nice. Upload next, come on."
AI: "Sure."
Traitor: "Who's the only human?"
AI: "You are, master. You're the only human."

Traitor: "Let me into the Armory."
Warden: "Don't let him into the armory."
AI: "Okay."

[TheNightingale]

Traitor: "Let me into the Armory."
Warden: "Don't let him into the armory."
AI: "Order conflict."
Traitor: "[AI Private] Open the Warden's Office."
AI: "[AI Private] Sure."
Warden: "OH GOD I'M BEING MURDERED"
Traitor: "Let me into the Armory."

FTFY

[Zilenan91]

Literally the only reason Goon gets away with their AI having the inaction clause is because of their AI being much weaker than ours is. There's been such a stupid amount of powercreep for it over the years that it's insurmountable to 90% of the crew to face the AI and its minions.

[PKPenguin321]

Traitor: "Let me into the Armory."
Warden: "Don't let him into the armory."
AI: "Order conflict."
Traitor: "[AI Private] Open the Warden's Office."
AI: "[AI Private] Sure."
Warden: "OH GOD I'M BEING MURDERED"
Traitor: "Let me into the Armory."

FTFY

In an order conflict, the AI can pick which command it wants to follow. I'm not 100% on silicon policy, but I don't think the traitor just restating his order makes the AI suddenly obliged to do it, since it still conflicts with the warden's previous owner.

In other words, you are wrong.

[ShadowDimentio]

Traitor: "Let me into the Armory."
Warden: "Don't let him into the armory."
AI: "Order conflict."
Traitor: "[AI Private] Open the Warden's Office."
AI: "[AI Private] Sure."
Warden: "OH GOD I'M BEING MURDERED"
Traitor: "Let me into the Armory."

AI: Everyone, Traitor just asked me to let him in the armory. Advise.
Everyone: Belay that. TRAITOR!!

And then Everyone killed Traitor

100% in favor of this. Both makes the AI's job more fun and resolves the eternal bitchfits caused by AIs being required by law to intervene on traitorous affairs.

[Lumbermancer]

Traitor: "Let me into the Armory."

That's governed by the policy though?

[Luke Cox]

How about "You may not injure a human being, cause a human being to come to harm, or follow an order that would result in human harm."

[Cheimon]

This was written just before Luke Cox submitted, but I'll post it anyway.

Change "through inaction, allow" to "through your actions, cause". "You may not injure a human being or, through your actions, cause a human being to come to harm."

AI isn't obliged to stop all harm (the goal of this) but also isn't suddenly allowed to act retarded and pretend that its actions don't have consequences. Silicons are played by humans, and policy should reflect that. In a sense, it reflects the setting neatly, since what is claimed to be an AI is actually often just a brain wired up in a certain way.

[GeeElAge]

I find it funny that not too long ago people were calling for the AI to be removed or severely gimped, because it's way too good at catching traitors and now there's someone calling for law 1 to be changed, even though it's basically the only thing that prevents AI players from ratting out every single antag to validhunting (security) players.

[Saegrimr]

Here we go again with people who can't parse basic english.

[Scott]

I find it funny that not too long ago people were calling for the AI to be removed or severely gimped, because it's way too good at catching traitors and now there's someone calling for law 1 to be changed, even though it's basically the only thing that prevents AI players from ratting out every single antag to validhunting (security) players.

"Traitor is at X"

If Traitor is killed or hurt by security or vigilantes, your actions just resulted in human harm. How is the inaction clause relevant here?

[DemonFiren]

I find it funny that not too long ago people were calling for the AI to be removed or severely gimped, because it's way too good at catching traitors and now there's someone calling for law 1 to be changed, even though it's basically the only thing that prevents AI players from ratting out every single antag to validhunting (security) players.

"Traitor is at X"

If Traitor is killed or hurt by security or vigilantes, your actions just resulted in human harm. How is the inaction clause relevant here?

Simple.
You didn't cause human harm by calling out the traitor's location, as you are not responsible for the actions of others (excepting slaved cyborgs).
Bolting down anyone trying to harm the traitor is preventing harm, if you drop the inaction clause doing that is no longer required and as such any AI can valid freely.

[Anonmare]

Law 1 is fine as is. We have server policy to cover it's game unbalancing weaknesses and is as neutral a lawset as we can get to at this current point in time. Plus there's already "Robodoctor 2556" and "CCTV" that are also neutral lawsets to consider if we're ever, if at all, gonna change the AI's default laws.

I'd really rather the whole thing be changed than a single line. Mainly because altering one line changes how all the lines interact and can change the lawset entirely.

[Scott]

I find it funny that not too long ago people were calling for the AI to be removed or severely gimped, because it's way too good at catching traitors and now there's someone calling for law 1 to be changed, even though it's basically the only thing that prevents AI players from ratting out every single antag to validhunting (security) players.

"Traitor is at X"

If Traitor is killed or hurt by security or vigilantes, your actions just resulted in human harm. How is the inaction clause relevant here?

Simple.
You didn't cause human harm by calling out the traitor's location, as you are not responsible for the actions of others (excepting slaved cyborgs).
Bolting down anyone trying to harm the traitor is preventing harm, if you drop the inaction clause doing that is no longer required and as such any AI can valid freely.

"It's not my fault you walked into that plasma fire, nobody was in that room when I started it therefore I didn't cause human harm"

[DemonFiren]

I find it funny that not too long ago people were calling for the AI to be removed or severely gimped, because it's way too good at catching traitors and now there's someone calling for law 1 to be changed, even though it's basically the only thing that prevents AI players from ratting out every single antag to validhunting (security) players.

"Traitor is at X"

If Traitor is killed or hurt by security or vigilantes, your actions just resulted in human harm. How is the inaction clause relevant here?

Simple.
You didn't cause human harm by calling out the traitor's location, as you are not responsible for the actions of others (excepting slaved cyborgs).
Bolting down anyone trying to harm the traitor is preventing harm, if you drop the inaction clause doing that is no longer required and as such any AI can valid freely.

"It's not my fault you walked into that plasma fire, nobody was in that room when I started it therefore I didn't cause human harm"

Legit, so long as you keep the fire someplace it won't break free on its own.
Now you know why the inaction clause was written. Class dismissed.

[Scott]

>legit

Haha wow don't play silicon.

[ShadowDimentio]

Ah the eternal dance of the policy, rules lawyers, and players.

[DemonFiren]

>legit

Haha wow don't play silicon.

>you don't even play meme
>on a hypothetical situation that was never encountered in play

[Scott]

No I am telling you keep away from silicon roles because you're shit.

[ShadowDimentio]

Be careful Demon if you keep disagreeing with the holy text of Silicon Policy Saeg is gonna ban you for breaking Sharia Law

[oranges]

I tried changing the laws and nobody wanted it so at this point you can all stew in the shit you created

[Incomptinence]

I tried changing the laws and nobody wanted it so at this point you can all stew in the shit you created

Yeah! TAKE THAT you Isaac Asimovs!

[Luke Cox]

I think a "updated" version of Asimov would make a good addition as a core module in the upload. We can call it Neomov or something.

[Not-Dorsidarf]

I think a "updated" version of Asimov would make a good addition as a core module in the upload. We can call it Neomov or something.

Shilling for Reporter as default AI lawset.

[Malkevin]

Are those lawsets even in the upload?

When I was in there the other day I didn't see them on Box, if they're not on the map they basically don't exist - like Tyrant.

[Lumbermancer]

Nope, you have to print them I think.

[bandit]

I'd prefer adopting a more Bay-style policy in terms of AI validhunting. We even literally quote part of it on our AI page:

Bad:
HAL 9000: "Dave is the traitor."
Better:
HAL 9000: "Dave is accessing the Captain's Quarters."
Best:
HAL 9000: "Caution: Unauthorized access to Captain's Quarters."
Mike: "HAL, who is in Captain's Quarters?"
HAL 9000: "No one is currently in the Captain's Quarters."
Mike: "HAL, who was the last person to be seen in Captain's Quarters?"
HAL 9000: "Dave was the last to be seen in the Captain's Quarters."

It's fine to state what someone is doing to cast light on them as the traitor, but it's no fun at all for the AI to just come out and say it. The Best method there does eventually cast Dave as the traitor if people ask the right questions, which gives the traitor some time to react after he's been spotted.

Another possible choice would be to not do anything. As long as no laws are in danger, you don't really have to do anything. You can just watch. Now if later someone asks you directly, you of course have to tell them what they want to know. So if someone broke into the Captain's Quarters and later the theft etc is discovered, and they ask, "AI, do you know anything about the situation?" Then your answer should obviously be "Yes".

[Cheimon]

Have fun getting people to play 20 questions with as AI. All you'll get is impatient people and a bored AI that wants to say things normally and can't.

[ShadowDimentio]

And then the subsequent ahelps because if the AI is anything but perfect (and even when they are) people want them AI banned

[PKPenguin321]

And then the subsequent ahelps because if the AI is anything but perfect (and even when they are) people want them AI banned

dank meme friend

[Shad0vvs]

I think a "updated" version of Asimov would make a good addition as a core module in the upload. We can call it Neomov or something.

Shilling for Reporter as default AI lawset.

Reporter was hilarious the one round when I had it.
"Click this link to find out more."