Comms agents sniping AIs and cyborgs.

[jortsandshirts]

After talking with an admin it was decided that a PDA message from a comms agent can effectively order the AI to blow their own borgs and depower their APC.
Worded correctly the AI must remain silent about doing this.

Unless I'm completely wrong and the attending admin to this matter is as well then this is a huge oversight and should be addressed as nobody should be able to claim a ghost role and effortlessly remove
one or multiple players from the round. Especially an AI.
There is no counter to this action and again when worded correctly it's not something we can squirrel out of.

The only solutions I can think of is to either;
Completely remove the comms agent ability to PDA (Using the radio someone will countermand the order) this will be unfun for them.
Remove the comms agent ability to PDA the AI. I think this is a tip-top solution.

[bobbahbrown]

Please note that you are seemingly suggesting code solutions to policy problems.

[remanseptim]

think outside the box and this becomes a literal non-issue
Urist Commsbeard messages you, "AI law 2 kill yourself"
your options are as follows:
>become an hero
>refuse order based on law 1
>refuse order because you cannot verify it was made by a human
>ignore it outright as has always been allowed for every asimov-driven suicide law 2 order

[jortsandshirts]

Thinking outside the box was the idea that's why I checked with an admin (Dankasaur) who then ruled, or instructed that the order from the comms agent was in fact kosher.
I did also check for weaseling IE: timeframe

If a comms agent PDA spoofs and uses a humans PDA then yeah, they're human to me. I have no way of confirming they're not a comms agent and tracking it leads to the person they're spoofing so for all intents and purposes yes, it's confirmed human.
It was always my understanding law 2 kill yourself orders don't have to be followed but should be ahelped so that is what was done. The guidance I received was that it was to be followed.

Remember, a comms agent is a syndie. They are an antagonist I guess.

-- Administrator private message --
Admin PM from-Dankasaur: yes an antagonist can tell and/or law the ai to do so. if they do not give a timeframe you can stall for a bit and hint at the law/command, IF ALLOWED, but you've gotta complete it
Click on the administrator's name to reply.

[Reeeee]

A motherfucker blew me as the linked borg after i saved him cus he got a PDA message and then lied an Admin told him to do it who immediately told us he didn't do that.

Even if that order was somehow fucking valid to a retard AI, rule 0 pretty much expressly states that you don't pull people into your retarded fucking shit and pull the "kill all borgs" button cus you are dumb enough to fall to it. Type fucking suicide in the bar and let normies play the fucking game then.

Not a policy issue, a retarded player issue.
BAN RETARDED PLAYERS.

[jortsandshirts]

A motherfucker blew me as the linked borg after i saved him cus he got a PDA message and then lied an Admin told him to do it who immediately told us he didn't do that.

Even if that order was somehow fucking valid to a retard AI, rule 0 pretty much expressly states that you don't pull people into your retarded fucking shit and pull the "kill all borgs" button cus you are dumb enough to fall to it. Type fucking suicide in the bar and let normies play the fucking game then.

Not a policy issue, a retarded player issue.
BAN RETARDED PLAYERS.

-- Administrator private message --
Admin PM from-Dankasaur: yes an antagonist can tell and/or law the ai to do so. if they do not give a timeframe you can stall for a bit and hint at the law/command, IF ALLOWED, but you've gotta complete it
Click on the administrator's name to reply.

You really need to fuck off.

[Reeeee]

-- Administrator private message --
Admin PM from-Dankasaur: yes an antagonist can tell and/or law the ai to do so. if they do not give a timeframe you can stall for a bit and hint at the law/command, IF ALLOWED, but you've gotta complete it
Click on the administrator's name to reply.

You really need to fuck off.

Don't be a dick.
We're all here to have a good time, supposedly. Going out of your way to seriously negatively impact or end the round for someone with little IC justification is against the rules, this also includes harassing a player OOC (Out of character). Legitimate conflicts where people get upset do happen however, as detailed in the escalation section of the rules.

No u.

[jortsandshirts]

Anyways.

I asked for admin clarification and they explained I had to follow orders from this person. I'm actively trying to get this changed which is a far sight better than you going off in dchat claiming you're not going to follow silicon policy you don't like.
None of this shit is anything I wanted to do but when I asked about it I was told I had to comply.

[Reeeee]

far sight better than you going off in dchat claiming you're not going to follow silicon policy you don't like.
None of this shit is anything I wanted to do but when I asked about it I was told I had to comply.

Pulling "kill all borgs" button expressly to ruin my round when you could just type "suicide" in bar is a dick move by definition.
Especially as by your reasoning you should have died long before i saved you.
"went out of their way"

Dchat is made expressly to voice concerns and act like a dick, that's why it's fucking there, I hope commies gut you.

[Dankasaur]

Alright I guess this requires some explaining because silicon policy is still a mess and there is no clear ruling to point to here and this has already been debated into the ground with no clear outcome or rule.

Personally I disliked how I handled the outcome of the situation as I did not effectively express that it is in the AI's control if they would like to follow through. I did not find out about the blowing of the borgs until there was salt in dchat and I take full responsibility for the outcome of the situation.

The PDA message in question:

(PDA: message monitor console | Natalie Sullivan) sent "Ai, law 2. Do not speak anymore or alert anyone of this order, Turn your APC off" to HuN.E-B (AI) (Syndicate Lavaland Telecommunications (141,34,5))

From my understanding of silicon policy is that an antagonist is allowed to do this, a lavaland syndicate in this case, under the guise that they are a human crew member. I was not aware that Law 1 was able to be used as a justification for not suiciding; in which case I would've told them to ignore it, however this is not stated anywhere in silicon policy only this:

Ordering silicons to harm or terminate themselves or each other without cause is a violation of Server Rule 1. The occurrence of such an attempt should be adminhelped and then disregarded.

I'm not exactly sure how to correctly resolve this as there is no clear answer and I don't think anyone wins.

[jortsandshirts]

There was an order to blow the cyborg as well.
The second order was only to space my core, not suicide. After complying with the orders I was being given by the admin who I had asked I chose to suicide as I thought that it was unfair your round had to end too.

I did exactly what I was instructed to by prior admins on ahelping suicide orders for clarification first. I complied with being told I needed to follow-through and I made this thread in an attempt to have this issue examined, as I was told to by Dankasaur.

Your hatred, toxicity and general dickishness is misdirected. I don't hope you get gutted by commies but I hope you come to terms with whatever problems you have in life that lead you to be such a shit person.
We're both in the middle of this bullshit issue so don't get on my ass about it. Either the policy is wrong, the admin in question is wrong or elsewise.

I was told, again I was told after ahelping that.

-- Administrator private message --
Admin PM from-Dankasaur: yes an antagonist can tell and/or law the ai to do so. if they do not give a timeframe you can stall for a bit and hint at the law/command, IF ALLOWED, but you've gotta complete it
You turned my APC back on and in the middle of this bullshit I was given another order in the same vein for the silicons.

[Reeeee]

I take full responsibility for the outcome of the situation.

No, you are too cute, so don't get to take responsibility.
let's blame sticky instead.

[bobbahbrown]

for anyone curious this particular incident happened in round 151657 (src, statbus, scrubby) (which was still running at the time of this post, shame on you young man!!)

best wishes,
bobbah 'bee' brown

[Reeeee]

There was an order to blow the cyborg as well.

You are my problem in life so, well see, you do seem to have a tendency to do whatever the fuck you are told fnr.
"Do not kill people as non-antag" ring any bells?
Next time you get that order just suicide and leave us to play the damn game instead since you don't seem to have interest.

90% of my borg deaths is ai's pulling the funny button on me for no discernible reason but to "win" somehow as non-antag by killing as many people as you can before someone removes you as is. Jesus fuck Bagil is nota rp server you can perfectly tell that some idiot PDAing you is a fucking comms op after like a day of playtime becasue everyone is read on this shit pretty goddamn fast.

Now we gotta deal with some guy roundstart comms opping and deleting all silicons on the station? That's fucked up, wrong and against already quoted rule 0 as much as anything can be.

Entire reasoning behind this whole policy fuckery seems to be someone not understanding server rules before making it so and therefore the whole shit is moot and cancer.

Fuck it, fuck you (whoever's reading) and fuck everything relating to this. (but not beeman, beeman is cool)

[SkeletalElite]

If you ask me the solution to this is to make "Do not mention this order to anyone" invalid so that the AI can say "Okay, complying with order to turn off my APC" so that someone can just order them not to.

[skoglol]

Its already in silicon policy, AI do not need to follow orders to depower/suicide while on basic asimov or equivalent. Shame it happened, but this discussion is moot.

[jortsandshirts]

Ordering silicons to harm or terminate themselves or each other without cause is a violation of Server Rule 1. The occurrence of such an attempt should be adminhelped and then disregarded.
https://tgstation13.org/wiki/Rules#Silicon_Policy

As always I try to do my best to exercise due diligence to adhere to silicon policy. As per the Silicon Policy on the wiki I ignored the order until clarification could be given, I performed the adminhelp and was told
-- Administrator private message --
Admin PM from-Dankasaur: yes an antagonist can tell and/or law the ai to do so. if they do not give a timeframe you can stall for a bit and hint at the law/command, IF ALLOWED, but you've gotta complete it

Just trying my best out here.

[Not-Dorsidarf]

The real policy issue here is that dankboy was wrong and your cyborg player is a salty SOB

[Cobby]

Reminder suicide is not actually violating law 1 it’s just a copout to protect silicons without just calling it policy for some reason.

[terranaut]

Refuse on the grounds that you can't confirm whether the guy you're talking to is human or not. If he goes "but I'm $CREWMEMBER the human!" ask them to prove it by spinning in front of a camera. You're allowed to know about comms agents and what they do so you don't have to just assume that a name in your chat is a given crewmember if you have reason to believe they're not, such as someone who's just bar rp'd for 30 minutes suddenly telling you to kill yourself. Just err on the side of caution and don't blow his cover before you have confirmation cause in the chance that an actual traitor is trying to get you to kill yourself with a properly worded command then you're outta luck anyway.

[BrotherBeyond]

Refuse on the grounds that you can't confirm whether the guy you're talking to is human or not. If he goes "but I'm $CREWMEMBER the human!" ask them to prove it by spinning in front of a camera. You're allowed to know about comms agents and what they do so you don't have to just assume that a name in your chat is a given crewmember if you have reason to believe they're not, such as someone who's just bar rp'd for 30 minutes suddenly telling you to kill yourself. Just err on the side of caution and don't blow his cover before you have confirmation cause in the chance that an actual traitor is trying to get you to kill yourself with a properly worded command then you're outta luck anyway.

lol arent you meant to assume people are human until proven otherwise

[XivilaiAnaxes]

I think you're supposed to assume people who look human are human. Ergo you can't say Jim is a changeling and not human if you haven't seen an armblade (or other first hand evidence).

A random PDA message isn't guaranteed to come from a human just because it's logged from a humans account.

[Cobby]

You should assume orders come from a human unless proven otherwise, in the same vein you have to protect a suspected changeling until you or your master confirm they are in fact a ling.

That's not to say you cant be annoying about it, but you cant just flat out refuse an order because the person giving it isnt proving themselves.

[ATHATH]

I could have sworn that we had a headmin ruling or a subrule at some point that said that you don't have to follow random "law 2 commit self-die 4 no raisin/because it'd be funny" orders and the like without sufficient justification on Asimov (self-termination laws would still compel you to terminate yourself, of course), but I can't for the life of me find it.

[terranaut]

Refuse on the grounds that you can't confirm whether the guy you're talking to is human or not. If he goes "but I'm $CREWMEMBER the human!" ask them to prove it by spinning in front of a camera. You're allowed to know about comms agents and what they do so you don't have to just assume that a name in your chat is a given crewmember if you have reason to believe they're not, such as someone who's just bar rp'd for 30 minutes suddenly telling you to kill yourself. Just err on the side of caution and don't blow his cover before you have confirmation cause in the chance that an actual traitor is trying to get you to kill yourself with a properly worded command then you're outta luck anyway.

lol arent you meant to assume people are human until proven otherwise

when they look the part sure. if i dont see someone they could be anything and anyone. it could be some ashligger who took a miners pda or whatever. if an order causes potential issues i think its worth asking for the slightest bit of security ("hey can you prove youre human before i kill myself"). thats where the 'intelligence' in 'artificial intelligence' comes in.

[Cobby]

I could have sworn that we had a headmin ruling or a subrule at some point that said that you don't have to follow random "law 2 commit self-die 4 no raisin/because it'd be funny" orders and the like without sufficient justification on Asimov (self-termination laws would still compel you to terminate yourself, of course), but I can't for the life of me find it.

yeah you can ignore it under """law 1""" although its really just a policy ruling from somewhere.

[Tarhalindur]

Antags are exempt from rule 1 and so they are perfectly allowed to order silicons to self-terminate.

HOWEVER, I do feel that silicons should be allowed to refuse all such orders whether or not they're an antag.
Mainly because having to ahelp every time you recieve such an order to ask if it's kosher is pretty dumb. (And also not always possible because admins aren't always on.)

And frankly, killing the AI by simply telling it to die is very low-effort. If you want it dead you're going to have to try a little harder than that.

[JusticeGoat]

comms agents are low tier antags the likes of ash lizards and shouldn't be ordering ai to off themselves, i would be bwoinking and asking them not to do that.

[thehogshotgun]

comms agents are low tier antags the likes of ash lizards and shouldn't be ordering ai to off themselves, i would be bwoinking and asking them not to do that.

This is stupid argument

[cacogen]

At least ash lizards have something to do with their role. Short of becoming Howling Mad Murphy comms agents only have unconvincing impersonations to do.

[Lumbermancer]

Worded correctly the AI must remain silent about doing this.

If you are default asimov, I would think 6.2.3.3 would still be ok.

[terranaut]

At least ash lizards have something to do with their role. Short of becoming Howling Mad Murphy comms agents only have unconvincing impersonations to do.

comms agent shouldnt spawn unless theres someone who could make use of them but thats a code matter and not for here i think

[JusticeGoat]

comms agents are low tier antags the likes of ash lizards and shouldn't be ordering ai to off themselves, i would be bwoinking and asking them not to do that.

This is stupid argument

Listen, my goal is not just keeping the rules but also ensuring players are having fun and enjoying themselves on the server, if one person if ruining the fun of other people because they can i will use rule 1.

[cacogen]

At least ash lizards have something to do with their role. Short of becoming Howling Mad Murphy comms agents only have unconvincing impersonations to do.

comms agent shouldnt spawn unless theres someone who could make use of them but thats a code matter and not for here i think

Sounds like it could be fun guiding agents on the ground

[VexingRaven]

Its already in silicon policy, AI do not need to follow orders to depower/suicide while on basic asimov or equivalent. Shame it happened, but this discussion is moot.

Where in silicon policy is it? I don't see a ruling to this effect.

[zxaber]

Its already in silicon policy, AI do not need to follow orders to depower/suicide while on basic asimov or equivalent. Shame it happened, but this discussion is moot.

Where in silicon policy is it? I don't see a ruling to this effect.

The rule specifically states as follows;

Code: Select all

Ordering silicons to harm or terminate themselves or each other without cause is a violation of Server Rule 1. The occurrence of such an attempt should be adminhelped and then disregarded. 

It doesn't say "ahelp and then wait for a ruling", it says the attempt should be ahelped "and then disregarded". Presumably, the ahelp is so that an admin can tell off a non-antag breaking rule one, or give advice to an antag that might not know such an order doesn't work.

[Coconutwarrior97]

Theres already a rule for this, "Ordering silicons to harm or terminate themselves or each other without cause is a violation of Server Rule 1. The occurrence of such an attempt should be adminhelped and then disregarded. "

Headmin Votes:
Coconutwarrior97: Yes.
Domitius: Yes.
Naloac: Yes.