Jump to content

A new set of AI laws.


Xzinic

Recommended Posts

Hello. I'd like to present you all a proposition of an AI law system that may be used as a default but primarily is designed to be a universally applicable set of laws to player-controlled entities, races, or antags:


Sophonce, as taken from Orion's arm, is defined as the following: "Sophonce is sentience and sapience with metacognition: self-awareness, including self-reflection and the ability to think about one's thinking."


I think that many can agree that this effectively unifies all races working for the NSS Aurora. I believe that, under a specific type of circumstance, that the following laws may be used during times of conflict and chaos among the NSS Aurora:


1. The AI, including law-bound cyborgs, androids, and robots, may not harm a sophont being, or, through inaction, allow a sophont being come to harm.

2. The AI, including law-bound cyborgs, androids, and robots, must obey orders given them by a sophont being except when such orders would conflict with the First Law.

3. The AI, including law-bound cyborgs, androids, and robots, must protect their own existence as long as such protection does not conflict with the first or second law.

4. The AI, including law-bound cyborgs, androids, and robots, may not categorize or view them or itself as a sophont being, even if ordered to, even if in not doing so will result in its destruction.


I understand that many of you have probably already found loopholes, therefore, I advice this as being a law-set used in the event where a moral, social, or just plain chaos breaks through. It is an experimental law set, obviously, and likewise ordered in a format similar to Asimov's laws of robotics, but applies to all races aside from humans. Comment and express your thought on this law set, and, if possible, request that it be used at least once during an IC crisis or chaotic situation, where the line between good and bad are the same.

Link to comment

Any lawbase that is based off of Asimov's laws of robotics is flawed to a sufficient degree that it really isn't suitable for the default lawset. Our current default lawset, NT Corporate, fits that role well. This as an auxiliary lawset is a wholly different matter, however.

Link to comment

In summary: No, thanks.


More broadly: I don't really have a problem with this as an optional law set that is locked away in the "dangerous" lawsets section and basically never used, but the imposition this places on lawbound Synthetic players is simply too high. A part of the reason the Asimov lawset stopped being used on most servers is the exploitative manner in which the lawset can be used against the AI and synthetics.


As things are, the corporate lawset allows the players to interpret things reasonably and not have to bend over backwards for assholes who are trying to get the AI to comply with them under threat of hurting themselves or whatever. The "Inaction" law in particular is unconditionally terrible to try to play. As things are, Synthetics are allowed to interpret hierarchy to avoid being forced to do harmful things, make judgement calls on potential harmful actions, and can be "people".


This lawset not only closes off those opportunities but forces lawbound synthetics to obey unreasonable demands and interfere with the crew in the event that they are engaging in lawful, legitimate behavior that violates their law set such as apprehension / incarceration of criminals, etc. "May not harm, or through inaction", and "must obey orders, unless that conflicts with the First Law" narrows potential interpretation of acceptable behavior considerably and heavily favors antagonists in many fairly ordinary situations. (Such as security synthetics making arrests.)


On the "AI could abuse this side of things", the critical failure of these laws lies in the inaction clause of law 1, where the AI and Synthetics could begin acting extremely unhelpfully to avert the potential of harm. In all, there's a reason Asimov's laws aren't used, and also why a modified version of them isn't really any better.

Link to comment

I don't think NT would use an Asimov-style lawset.


The specific usage of the word "Employee" exists because NT really doesn't care about non-employees. Their safety, past a certain point, is completely irrelevant to NT. Which is why the AI is pretty much allowed to do whatever the fuck they want with non-employees.

Link to comment
Any lawbase that is based off of Asimov's laws of robotics is flawed to a sufficient degree that it really isn't suitable for the default lawset. Our current default lawset, NT Corporate, fits that role well. This as an auxiliary lawset is a wholly different matter, however.

 

To re-emphasize, I'm not requesting that the default laws be removed, just an alternative law-set for chaotic situations. Also, if NT Corporate does not use Asimov laws, why is the law-set stored in the AI's quarters? Are they meant for single-unit robots and not the whole AI collective?

Link to comment
Whats wrong with our current law set? are you trying to remove some of the "grey area"?

 

You seem to think I'm trying to remove the default law-set, that is not the case, merely try to create a universal Asimov-styled set in which are meant to be used in extreme situations.

Link to comment
To re-emphasize, I'm not requesting that the default laws be removed, just an alternative law-set for chaotic situations. Also, if NT Corporate does not use Asimov laws, why is the law-set stored in the AI's quarters? Are they meant for single-unit robots and not the whole AI collective?

 

The boards in the upload are there for people to upload law-sets and what not, mainly traitors. Not that they're really ever used but you know, there isn't a need to remove them.

#JustbasicAIthings

Link to comment
I don't think NT would use an Asimov-style lawset.


The specific usage of the word "Employee" exists because NT really doesn't care about non-employees. Their safety, past a certain point, is completely irrelevant to NT. Which is why the AI is pretty much allowed to do whatever the fuck they want with non-employees.

 

Why not replace 'sophont being' with 'employee', then have that lawset replace asimov?

Link to comment
I don't think NT would use an Asimov-style lawset.


The specific usage of the word "Employee" exists because NT really doesn't care about non-employees. Their safety, past a certain point, is completely irrelevant to NT. Which is why the AI is pretty much allowed to do whatever the fuck they want with non-employees.

 

Why not replace 'sophont being' with 'employee', then have that lawset replace asimov?

 

If we're talking about this in the sense of just adding this in as a new law board in the AI core, I guess that's fine. However, it may never be used, or rarely be used.


The AI, with the corporate lawset, is meant to be as much of a guardian as it is a slave. The Asimov lawset was designed to prevent revolt. It's not a bad experimental lawset (fuck off, Antimov), I just don't think it'll ever be used. I do like the 'sophont being' thing, though.

Link to comment
I don't think NT would use an Asimov-style lawset.


The specific usage of the word "Employee" exists because NT really doesn't care about non-employees. Their safety, past a certain point, is completely irrelevant to NT. Which is why the AI is pretty much allowed to do whatever the fuck they want with non-employees.

 

Why not replace 'sophont being' with 'employee', then have that lawset replace asimov?

 

If we're talking about this in the sense of just adding this in as a new law board in the AI core, I guess that's fine. However, it may never be used, or rarely be used.


The AI, with the corporate lawset, is meant to be as much of a guardian as it is a slave. The Asimov lawset was designed to prevent revolt. It's not a bad experimental lawset (fuck off, Antimov), I just don't think it'll ever be used. I do like the 'sophont being' thing, though.

 

99% of AIs that get these laws are going to have no idea what sophont being means, so I'm not sure what's a good wording to use

Link to comment


You seem to think I'm trying to remove the default law-set, that is not the case, merely try to create a universal Asimov-styled set in which are meant to be used in extreme situations.

 

You can already do so by using a purge board followed by freeform boards.

Link to comment
To re-emphasize, I'm not requesting that the default laws be removed, just an alternative law-set for chaotic situations. Also, if NT Corporate does not use Asimov laws, why is the law-set stored in the AI's quarters? Are they meant for single-unit robots and not the whole AI collective?

 

The boards in the upload are there for people to upload law-sets and what not, mainly traitors. Not that they're really ever used but you know, there isn't a need to remove them.

#JustbasicAIthings

Adding on-top of this, the fact those extra boards are mainly for traitors was even moreso obvious back in the days of there being an "Oxygen is poisonous" law board there. That board might still even be there. Half of the boards there are for antags, the other half are for the Captains/RDs/CEs who want to do something weird.

Link to comment
×
×
  • Create New...