Silicon Protections 2.1.6.5 - interpretation and concerns

charmisokay
Joined: Tue Jul 24, 2018 9:44 am
Byond Username: Charmisokay

Silicon Protections 2.1.6.5 - interpretation and concerns

Post by charmisokay » #456763

Bottom post of the previous page:

To me, rule 2.1.6.5 always seemed very clear. Other players voiced their interpretations of the rule and they were very different from mine (and often each other's, as well).

I think the rule should be rewritten so it's very clear what it's supposed to state.

The most relevant rules:
-2.1.6: Any silicon under Asimov can deny orders to allow access to the upload at any time under Law 1 given probable cause to believe that human harm is the intent of the person giving the order.
-2.1.6.1: Probable cause includes presence of confirmed traitors, cultists/tomes, nuclear operatives, or any other human acting against the station in general; the person not having upload access for their job; the presence of blood or an openly carried lethal-capable or lethal-only weapon on the requester; or anything else beyond [metagamiing] that indicates the person seeking access intends redefinition of humans that would impede likelihood of or ability to follow current laws as-written.
-2.1.6.3: You are allowed, but not obligated, to deny upload access given probable cause.
-2.1.6.5: In the absence of probable cause, you can still demand someone seeking upload access be accompanied by another trustworthy human or a cyborg.

Taken together, I understood rule 2.1.6.5 to be a safety measure that the AI could always use. (well, within reason as per rule 1. don't be a dick. More on this later).

I've heard two other main interpretations of rule 2.1.6.5. I've also heard a few clear misunderstandings of the rule, including an admin that used the following example (paraphrased): "Rule 2.1.6.5 exists so that assistants etc can't just rush up and demand entry" - missing the fact that rule 2.1.6.1 already gives the AI probable cause to deny assistants entry since they don't have access to the upload by default (marked with bold, above).

Onto the 'main' interpretations (I hope I understood their interpretations properly):

1) "2.1.6.5 can only be used against the captain / HoP / RD if you have reasonable suspicion that human harm is the intent of that person"
I don't think this is reasonable since such suspicion would, in any case I can imagine, be based on something that would result in probable cause anyways (and often fall under this clause: "anything else beyond [metagamiing] that indicates the person seeking access intends redefinition of humans that would impede likelihood of or ability to follow current laws as-written.") Besides, rule 2.1.6.5 implies that the person seeking entry should be trustworthy (by use of the phrase 'accompanied by ANOTHER trustworthy person..."). If you have some sort of suspicion that the person seeking access will cause harm, they're not really trustworthy now are they?
Finally, nowhere does the ruling imply that you need any form of suspicion. If this is the way the rule is meant to be understood I think it should be rewritten as thus:
"2.1.6.5: In the absence of probable cause, you can still demand someone seeking upload access be accompanied by a trustworthy human or a cyborg if you have a reason to suspect that the person seeking upload access might intend human harm. This suspicion must be based on something that has occured in the round." - However, I once again stress that I have a hard time imagining a suspicion that would not be covered by the probable cause clause.

2) "2.1.6.5 is a gameplay suggestion meant to give the AI a responsible way to let people into the AI upload despite having probable cause to deny them"
I don't think this is a reasonable interpretation either, because the rule's phrasing does not imply that 2.1.6.5 is merely a suggestion. I do, however, agree that asking for an escort is a very reasonable and responsible way to grant somebody access.
If the rule should be understood in this way, I think it should be rewritten as thus:
"In a case where the AI has probable cause, but is still unsure if denying the person entry is the right move, an ideal way of handling the situation is demanding that the requester is accompanied into the upload by another trustworthy human or cyborg" - But why make this a rule at all? Adding a gameplay suggestion as a rule seems very weird to me. Maybe forfeit the numbers (that indicate that this is a rule) and instead insert the suggestion in quotation marks at the end of the section on rule 2.1.6.
-an admin voicing this interpretation used the example (again paraphrased) of an RD seeking entry to the AI upload, without having access to it at round start (we later found out that RDs do have access, but that is besides the point). The RD is a trusted head of staff, but they don't have access to the AI upload (in this example). Therefore, the AI has probable cause but instead of denying them access a responsible way of letting them into the upload is demanding a trustworthy escort.

I would like to once again defend why I think the original ruling as I interpret it should be enforced, as well as a suggestion on how to make it even more clear that this is the intended way to interpret the rule (if that is necessary. I don't think it is).

-Rule 2.1.6 and rule 2.1.3 pretty much say "If you have reasonable cause to suspect that they might upload a dangerous law, you can deny them access or let them in at your discretion" (rule 2.1.6.4 however says that if they've caused human harm you must keep them out of the upload). Rule 2.1.6.5 then says: "Even if you don't have reasonable suspicion, you can still demand someone seeking upload access be accompanied by another trustworthy human or a cyborg." It does not imply anywhere that 'specific circumstances' must be met. A lot of people say 'but the captain has access so you must let him in', however, the rule is only applicable to people that naturally have access to the AI upload area. Therefore, if you can not use the rule against 'innocent' captains etc., it's a redundant rule that can never be applied. Under probable cause you can already let them in or deny them entry at your own discretion.

-Asking another trustworthy human to join you requires nearly no effort and takes very little time. I've always made sure to check the crew manifest and the scanner to see if there's a plethora of suited candidates available. When there's ~3 members of security and ~3 other heads of staff, one of them will answer. I've even let people into the upload alone if nobody responded to the call (since it would be a dick move to keep them out simply because everybody else is busy or deaf).

-On the other hand, changing the AIs laws have repercussions for the AI and potentially for the crew for potentially the rest of the shift (if the law is shitty or malicious). Of course the laws might also introduce an element of FUN, but the system I'm defending does not prevent law uploads. More on that below.

-This is a very reasonable safety measure from an IC and OOC point of view. It's not very intrusive and if the second person is not a memelord or an enemy of the station it will pretty much guarantee that no bad laws will be uploaded (poor wording and malicious laws both included). Also there's a very easy counterplay. You only need one other player to assist you. Captains and HoS' are well known for being in cahoots when it comes to anti greytide laws. The captain and HoP often agree on meme laws. However, if you're both baddies other people might move to oppose you. I've had a lot of RDs save my digital butt from horrible captain laws. Role interactions and conflicts make the game more interesting, yeah?
Speaking of counterplay, there's a spare AI upload board in the secure tech storage, and they can somewhat easily be printed via RnD.

-For the reasons outlined above, if it's within the AI's right to ask for an escort even without probable cause, I don't think it's a dick move (and thus against rule 1) unless you're demanding an impossible escort or refusing to listen to arguments. Usually when I've asked this, the person will either comply and we're done in ~2 minutes or they scream that I must obey them or that they're going to destroy me, after which they'll usually try to break into the upload and either manage to do it, or give up.

As for how the ruling should be phrased, I personally think the rule right now is already very clear, but if it should be made clearer following the interpretation I'm defending, it might look a little like this:
"2.1.6.5. You can always demand someone seeking upload access be escorted by another trustworthy human or a cyborg, even without probable cause. However, if nobody steps up to escort the person seeking upload access after a request for an escort has been made, you may not deny them access unless you have probable cause. Consult rule 1 (don't be a dick)."

"2.1.6.6. Suitable escorts include human heads of staff and human members of the security team that have not given indication that they will cause human harm. If no such person is available, you may not deny the person seeking upload access unless you have probable cause."


I think this is a decent wording since it reminds the AI that it can't just stonewall tell people to fuck off without probable cause, but it also protects the AI from the captain, HoP or RD uploading a malicious or meme law without any input from anyone else.
Last edited by charmisokay on Sat Nov 24, 2018 11:32 pm, edited 4 times in total.
Actionb
Joined: Thu Feb 05, 2015 8:51 am
Byond Username: Actionb

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Actionb » #457557

Lumbermancer wrote:
Actionb wrote:My point is: no OOC policy must override IC AI law.
Then you must write, or use another lawset, because Asimov is exploitable by design.
Exactly my point. Fuck I wish my brain would shut up sometimes so I can express the essence of my agenda.
User avatar
Lumbermancer
Joined: Fri Jul 25, 2014 3:40 am
Byond Username: Lumbermancer

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Lumbermancer » #457560

Is ok, I'm alt tabing from work so I may not comprehend fast enough.

But the problem with writing new lawset is, that if you want for it to be autoritative and/or subservient, it will always be exploitable. Unless of course you gonna make it really detailed and complex, but then what's the point, that's exactly what policy does right now.
aka Schlomo Gaskin aka Guru Meditation aka Copyright Alright aka Topkek McHonk aka Le Rouge
Image
somerandomguy
Joined: Sun Nov 05, 2017 7:41 pm
Byond Username: Astatineguy12

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by somerandomguy » #457568

>can't make an excuse for silicon policy
hahahahahhahaha
Just say you're hardcoded to interpret a lack of upload access as likely harm

>OOC policy shouldn't override laws
Well good luck with either AIs that lock everything down to stop harm or lawsets that you need a law degree to understand
User avatar
WarbossLincoln
Joined: Wed Feb 10, 2016 11:14 pm
Byond Username: WarbossLincoln

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by WarbossLincoln » #457582

After some time in the other thread on this I propose we delete the silicon policy page and replace it with:

1: follow server rules
2: follow your laws in priority order

job ban people who consistently fuck it up. It really doesn't need to be more complicated
--Crocodillo

Image
User avatar
Grazyn
Joined: Tue Nov 04, 2014 11:01 am
Byond Username: Grazyn

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Grazyn » #457583

"AI kill all lizards"

The only reason why an asimov AI doesn't obey that order is policy.

Asimov is flawed because it is a literary device meant to create conflict. But the madmen who run this server don't want it to create conflict, so policy is needed. So policy always comes first and at this point I think they're aiming to cover every single possible source of conflict because as goof said it's insane to have 5 subsections of a rule
User avatar
Anonmare
Joined: Sun Mar 15, 2015 8:59 pm
Byond Username: Anonmare

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Anonmare » #457606

Actually the AI can, and probably should follow that order.

An AI just can't do it on their own without prompting and they're also free to be as loud and roundabout or slow in the way they do it. They could also just kill wags-his-tail and say they killed all the lizards as ordered.
Image
Image
Image
User avatar
Lumbermancer
Joined: Fri Jul 25, 2014 3:40 am
Byond Username: Lumbermancer

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Lumbermancer » #457625

bwoink

Obviously unreasonable or obnoxious orders (collect all X, do Y meaningless task) are a violation of Server Rule 1. The occurrence of such an attempt should be adminhelped and then disregarded.
Spoiler:
oy vey unless you're an antag
aka Schlomo Gaskin aka Guru Meditation aka Copyright Alright aka Topkek McHonk aka Le Rouge
Image
User avatar
TehSteveo
Joined: Sat Apr 19, 2014 1:31 am
Byond Username: TehSteveo
Location: Pennsylvania

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by TehSteveo » #457632

Silicon policy is something nobody wants to touch as it's a major headache. It really does boil down to Asimov being the lawset as Asimov's three laws are ambiguous to begin with coupled with the fact it's a human roleplaying behind that silicon. We would probably be better changing the default lawset to something else, such as corporate lawset which makes sense for Nanotrasen to run. The issue there is, well, now you have a powergamey lawset as default. That or change to a modified Asimov lawset.
Freedom
Dr_bee
Joined: Fri Dec 23, 2016 6:31 pm
Byond Username: DrBee

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Dr_bee » #457635

TehSteveo wrote:Silicon policy is something nobody wants to touch as it's a major headache. It really does boil down to Asimov being the lawset as Asimov's three laws are ambiguous to begin with coupled with the fact it's a human roleplaying behind that silicon. We would probably be better changing the default lawset to something else, such as corporate lawset which makes sense for Nanotrasen to run. The issue there is, well, now you have a powergamey lawset as default. That or change to a modified Asimov lawset.
Or just let borgs powergame like every other player does. Why exactly are borgs expected not to powergame on the most powergame heavy server SS13 probably has is beyond me.
Cik
Joined: Thu Oct 30, 2014 2:24 pm
Byond Username: Cik

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Cik » #457637

corporate really sucks though, for numerous reasons besides powergame

look, asimov is the default because
1. the fluff is cool
2. it makes the AI a (relatively) neutral arbiter, that does not have an unreasonable hateboner for either the crew or the antagoniste(s) so that the AI does not use it's (increasingly less formidable) power to fuck everyone else minute 1 second 1 of every round for all eternity.

originally, the AI's job was basically:
1. help smooth out the access system
2. reduce greytiding via 1
3. prevent minute 1 instaunstoppable murderboner by having relatively robust servants and good intelligence / control capability
4. be an antagoniste (occasionally)

asimov is a good lawset. there is no reason to change it. AI policy is also fine, if a little wordy (the thrust is good, but there are too many words as it tries to handle every single little case that's ever come up ingame)

literally 99.9% of the time you just follow your laws in order and it's literal all gucci homie, these little edge cases come up once in a hundred rounds and if you ahelp it they will tell you what to do to not get b&

simple stuff
User avatar
TehSteveo
Joined: Sat Apr 19, 2014 1:31 am
Byond Username: TehSteveo
Location: Pennsylvania

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by TehSteveo » #457640

Dr_bee wrote:
TehSteveo wrote:Silicon policy is something nobody wants to touch as it's a major headache. It really does boil down to Asimov being the lawset as Asimov's three laws are ambiguous to begin with coupled with the fact it's a human roleplaying behind that silicon. We would probably be better changing the default lawset to something else, such as corporate lawset which makes sense for Nanotrasen to run. The issue there is, well, now you have a powergamey lawset as default. That or change to a modified Asimov lawset.
Or just let borgs powergame like every other player does. Why exactly are borgs expected not to powergame on the most powergame heavy server SS13 probably has is beyond me.
Somehow you're implying the AI/Borgs don't already powergame within their confines. As to why have barriers is the fact they are essentially multiple humanoid professions combined into their module with essentially all-access, ability to deny access with a ctrl+click, and immunities to things that would kill or hinder others. There has to be a downside and laws can serve as a downside, as well as a way to have somewhat more diverse gameplay occurring that requires thoughts such as command needing to change AI laws to deal with threats.
Freedom
User avatar
TehSteveo
Joined: Sat Apr 19, 2014 1:31 am
Byond Username: TehSteveo
Location: Pennsylvania

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by TehSteveo » #457641

Cik wrote:corporate really sucks though, for numerous reasons besides powergame

look, asimov is the default because
1. the fluff is cool
2. it makes the AI a (relatively) neutral arbiter, that does not have an unreasonable hateboner for either the crew or the antagoniste(s) so that the AI does not use it's (increasingly less formidable) power to fuck everyone else minute 1 second 1 of every round for all eternity.

originally, the AI's job was basically:
1. help smooth out the access system
2. reduce greytiding via 1
3. prevent minute 1 instaunstoppable murderboner by having relatively robust servants and good intelligence / control capability
4. be an antagoniste (occasionally)

asimov is a good lawset. there is no reason to change it. AI policy is also fine, if a little wordy (the thrust is good, but there are too many words as it tries to handle every single little case that's ever come up ingame)

literally 99.9% of the time you just follow your laws in order and it's literal all gucci homie, these little edge cases come up once in a hundred rounds and if you ahelp it they will tell you what to do to not get b&

simple stuff
Asimov is a good lawset, but it does need some precedents that get wordy. Asimov could be redone to better help reduce some confusion/bloat. As far as corporate, I disagree to an extent. I feel the law is much more exploitable to law changes, but I feel it is neutral in someways. Everyone interrupts it differently though which allows for the wiggle room for the lawset to shit on antagonist, even though technically, that traitor is a member of the crew and is expensive.
Freedom
Cik
Joined: Thu Oct 30, 2014 2:24 pm
Byond Username: Cik

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Cik » #457648

it has been a long while since i've analyzed it critically, but i remember being able to bend it over backwards to allow some really insane shit way back.

asimov is, in the most important ways, ironclad. like sure, without sillicon policy you can murder aliens but that's the only thing you can do that will make a ton of damage minute 1 second 1, and even then you can't do it in a way that has even a relatively low chance of harming humans, so no plasma floods, engine breaches, or shocking every station door etc.

most non-asimov lawsets either

1. bias you in one direction in the crew <----> antag axis or
2. take the chains off your ability to use maximal force in executing whatever your mandate is or
3. both

since this tends to ruin games for potentially lots of people, roundstart paladin/robocop/corporate/whatever is widely derided (or was, i suppose) and asimov is default lawset.
User avatar
oranges
Code Maintainer
Joined: Tue Apr 15, 2014 9:16 pm
Byond Username: Optimumtact
Github Username: optimumtact
Location: #CHATSHITGETBANGED

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by oranges » #457651

why is it so hard to just obey the server rules while also trying to RP the lawset to the best of your ability?
User avatar
Anonmare
Joined: Sun Mar 15, 2015 8:59 pm
Byond Username: Anonmare

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Anonmare » #457660

oranges wrote:why is it so hard to just obey the server rules while also trying to RP the lawset to the best of your ability?
It all went downhill after Saeg left

Prove me wrong
Image
Image
Image
User avatar
Lumbermancer
Joined: Fri Jul 25, 2014 3:40 am
Byond Username: Lumbermancer

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Lumbermancer » #457664

Image
aka Schlomo Gaskin aka Guru Meditation aka Copyright Alright aka Topkek McHonk aka Le Rouge
Image
User avatar
Cobby
Code Maintainer
Joined: Sat Apr 19, 2014 7:19 pm
Byond Username: ExcessiveUseOfCobby
Github Username: ExcessiveUseOfCobblestone

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Cobby » #457670

Malkraz wrote:
Cobby wrote:Again, it's not hard to interpret, it's simply admins not accepting the rule and should be made into a complaint.
There's a headmin ruling on record listed on the wiki stating the opposite and another for a similar, more general matter. That doesn't really sound to me like "simply admins not accepting the rule".
Complaint the headmin then, it's word for word allowed by silicon policy. Unless you outright misread it in the context of it's location, it's extremely straightforward.

If a headmin wants to enforce rules that go against the rule list they're freely allowed to change to make the server a better place in their eyes, I'd highly recommend for said headmin to review why they're in the position.
Voted best trap in /tg/ 2014-current
User avatar
Grazyn
Joined: Tue Nov 04, 2014 11:01 am
Byond Username: Grazyn

Re: Silicon Protections 2.1.6.5 - interpretation and concerns

Post by Grazyn » #457724

Anyone who remembers XISC can tell how FUN it was to play without silicon policy
Locked

Who is online

Users browsing this forum: No registered users