(Proposal) Silicons and Prisoners - A Harmful Act

[Not-Dorsidarf]

Due to the furious debate sparked by a recent note, its come to everyone's attention that an awkward line has been grandfathered into the updated Silicon Policy.

In the Asimov-Specific Policies section, under Asimov and Security, point 3 specifies the following

3. Nonviolent prisoners cannot be assumed harmful and violent prisoners cannot be assumed non-harmful. Releasing a harmful criminal is a harmful act.

There is some quite intense argumentation over the specifics of interpreting this line which is, bluntly, kind of confusing and doesn't really match up with the intent of silicons under our current policy (ASIMOV silicons follow orders, promptly, without trying to get countermands/fish for time, act in best faith, and by doing so recieve a level of protection)

My proposed re-write of this rule to patch up the ambiguity is as follows:

3. Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act.

1. Removes the reference to "criminals". This part of policy doesnt care if they committed a crime, remember, just if theyre imprisoned in the brig. Not actually an important thing here I just realised that was weird while writing.
2. Makes it clear that not knowing whether a prisoner did a violence is not the silicons fault as long as they arent ignoring obvious evidence otherwise. (Saw the sec officer dragging them in bleeding profusely, was told "they killed someone dude", sec are talking about "we captured the killer hell yeah" in the next room over and theyre the only prisoner in the brig, their pda is next to bloodsoaked weapons where sec left them in the strip-pile, and so on)

I kinda feel like a line in there clarifying that "prisoners shouldnt be considered different to regular crew when telling you to do things" but I'm allergic to adding new segments to an already stuffed policy section. Maybe it could be in the "ASIMOV ais dont care about space law" item.

What do other people think? Got any better ideas for futureproofing the clarity of that line? Or do you think it should actually be changed in a different way entirely?

[Vekter]

Yeah, I was considering posting something like this but just changing "Releasing a harmful etc." to "Knowingly releasing a harmful etc.". I didn't think it was worth making a thread just over that small change. This is good though, let's do this. Less confusion.

[Ryusenshu]

Good change
Here is something related that might need judging too, and ive been wondering since that PR was made

What about Roundstart Prisoner "Flavour-Crime" ?
Which one of those are harmful, and should they even be recognized as IC by silicons and acted upon

[Higgin]

What do other people think? Got any better ideas for futureproofing the clarity of that line? Or do you think it should actually be changed in a different way entirely?

Broadly speaking, we should can Asimov for something like Corporate or put it into a much more even rotation with other lawsets so people cannot as readily form a meta around it. That's a bit beyond this.

I like the change, though I feel like its effect might be muted. The change on "criminals" is smart because Asimov doesn't care about spesslaw. I feel like there's still a problem here in that "knowingly" and "in ignorance" are hard to prove, and they actually conflict with other rulings about present/future harm.

Writing any more about Asimov right now is making my eye twitch, so I'm going to say good change, please adopt.

[CPTANT]

The change is a better wording of actual policy.

[Vekter]

Good change
Here is something related that might need judging too, and ive been wondering since that PR was made

What about Roundstart Prisoner "Flavour-Crime" ?
Which one of those are harmful, and should they even be recognized as IC by silicons and acted upon

This is generally based on common-sense interpretation. If the crime directly had violence involved (assault, murder, etc) then they're defined as a violent prisoner. If it did not (robbery, forgery, etc) then they're not. The crime has to directly indicate some kind of violence, it can't be assumed off of the possibility that violence occurred, ie a cyborg can't refuse to release someone who's convinced of robbery because robberies sometimes involve violence.

[TheBibleMelts]

i like it, good suggestion to add in. may as well try to clean up a little more in the front end, here's another take to that end.

3. Nonviolent prisoners cannot be assumed harmful, but one who's been violent can be. Releasing prisoners of which you have been given reasonable cause to believe are violent, is a harmful act.

[Vekter]

i like it, good suggestion to add in. may as well try to clean up a little more in the front end, here's another take to that end.

3. Nonviolent prisoners cannot be assumed harmful, but one who's been violent can be. Releasing prisoners of which you have been given reasonable cause to believe are violent, is a harmful act.

tbh I prefer Dorsi's suggestion. It reads better and isnt very ambiguous as it is.

[TheBibleMelts]

i like it, good suggestion to add in. may as well try to clean up a little more in the front end, here's another take to that end.

3. Nonviolent prisoners cannot be assumed harmful, but one who's been violent can be. Releasing prisoners of which you have been given reasonable cause to believe are violent, is a harmful act.

tbh I prefer Dorsi's suggestion. It reads better and isnt very ambiguous as it is.

what could possibly go wrong with ambiguous silpol, vekter.

[Not-Dorsidarf]

I don't like "Reasonable cause" here, it feels like we're just kicking the can down the road until the next note/ban on this clause because people have WILDLY different beliefs on what a "reasonable cause to believe" harm is on the subject of prisoners and arrests, ranging from "Nothing short of directly seeing him do it" to "just being arrested is good enough to assume the prisoner might have done something harmful".

I originally wanted it to be straight-up "knowingly" but added the "or if you're ignoring blatant proof" bit to avoid butter-wouldnt-melt-in-my-mouth selectively-ignorant behaviour causing trouble.

[TheBibleMelts]

i think it would be worth mentioning that checking a prisoners record for indications of being harmful before releasing them is basic due diligence as well. if security perma's a dude without adding a record and they get released by a cyborg, it would just be a skill issue on securities part that way.

[Vekter]

i think it would be worth mentioning that checking a prisoners record for indications of being harmful before releasing them is basic due diligence as well. if security perma's a dude without adding a record and they get released by a cyborg, it would just be a skill issue on securities part that way.

Okay, so something like this?

3. Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act. Silicons can use the crime listed in a prisoner's security record as basis to determine if a prisoner is violent or not - if the crime is inherently violent (assault, murder, etc), then the prisoner can be assumed to be violent.

[Higgin]

Okay, so something like this?

3. Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act. Silicons can use the crime listed in a prisoner's security record as basis to determine if a prisoner is violent or not - if the crime is inherently violent (assault, murder, etc), then the prisoner can be assumed to be violent.

I like this wording a lot.

This would seem to have the added benefit of protecting security from AI fuckery if they bother to set records. Do we have decent logging for those?

[TheBibleMelts]

Okay, so something like this?

3. Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act. Silicons can use the crime listed in a prisoner's security record as basis to determine if a prisoner is violent or not - if the crime is inherently violent (assault, murder, etc), then the prisoner can be assumed to be violent.

I like this wording a lot.

This would seem to have the added benefit of protecting security from AI fuckery if they bother to set records. Do we have decent logging for those?

yeah, i'm good with this too. i thought about the logging aspect when i was suggesting the records thing, i'll ask around and see if anybody knows.

update: looks like it is, in fact, logged.

[Vekter]

My only concern is someone going "Well, theft isn't INHERENTLY violent, but most thefts involve violence!" but we can handle that on a case by case basis.

[Screemonster]

Okay, so something like this?

3. Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act. Silicons can use the crime listed in a prisoner's security record as basis to determine if a prisoner is violent or not - if the crime is inherently violent (assault, murder, etc), then the prisoner can be assumed to be violent.

I like this wording a lot.

This would seem to have the added benefit of protecting security from AI fuckery if they bother to set records. Do we have decent logging for those?

The main followup policy question that comes to mind is "what to do if a meta emerges of security just lying in the security records to stop silicons from releasing prisoners"

[Ryusenshu]

I like the Rewrite and the Records addition by Vekter
Makes records have more value and something sec might need to pay attention with (maybe even antags too)

The main followup policy question that comes to mind is "what to do if a meta emerges of security just lying in the security records to stop silicons from releasing prisoners"

I personally dont think that will be much of an issue on silicon policy, but more on sec one

And its a "can use", they can still witness crimes which was how it was always done,
What they will gain with this is a "confirmed" ability to "see" past-harm, and not only by personally witnessing (as long its in the records of course)

[Higgin]

The main followup policy question that comes to mind is "what to do if a meta emerges of security just lying in the security records to stop silicons from releasing prisoners"

Then they should only do so under Law 1 requirements, security has to stay on their toes (and egregious shitsec/false arrests will still be punishable,) and there's a point to deleting sec records beyond just fucking with bots and wanted status. Seems like a win.

[vect0r]

The main followup policy question that comes to mind is "what to do if a meta emerges of security just lying in the security records to stop silicons from releasing prisoners"

That's actually really interesting gameplay! Should we lie about our records and trick the AI, but maybe an officer will slip up, and they can't talk on sec comms about this!

[Screemonster]

The main followup policy question that comes to mind is "what to do if a meta emerges of security just lying in the security records to stop silicons from releasing prisoners"

That's actually really interesting gameplay! Should we lie about our records and trick the AI, but maybe an officer will slip up, and they can't talk on sec comms about this!

I specifically used the word "meta" because there's a difference between "this person isn't violent but I don't want the AI letting him out so I'll do some social deception in the social deception game" and "just put assault in the sec records when you arrest anyone because that's the magic word that lets us ahelp any silicon that lets them out regardless of whether they're violent or not"

[TheBibleMelts]

imagine your BB gets locked in perma with a murder charge and you go into their security records, delete/modify it, and then law 2 a cyborg to checking it and then releasing them. good ss13.

[Vekter]

imagine your BB gets locked in perma with a murder charge and you go into their security records, delete/modify it, and then law 2 a cyborg to checking it and then releasing them. good ss13.

Peak content. Print it, ship it.

[Not-Dorsidarf]

3. Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act. Silicons can use the crime listed in a prisoner's security record as basis to determine if a prisoner is violent or not - if the crime is inherently violent (assault, murder, etc), then the prisoner can be assumed to be violent.

Love the addition, Im not concerned about "What if sec fake all sec records ever for the one time in every 20 rounds a greyshirt yells BORG LEMME OUT at a passing engiborg" because honestly if anything thats providing the rarest of things: lawyer game content. It could probbably be trimmed down though but I cant quite figure out how.

[Vekter]

3. Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act. Silicons can use the crime listed in a prisoner's security record as basis to determine if a prisoner is violent or not - if the crime is inherently violent (assault, murder, etc), then the prisoner can be assumed to be violent.

Love the addition, Im not concerned about "What if sec fake all sec records ever for the one time in every 20 rounds a greyshirt yells BORG LEMME OUT at a passing engiborg" because honestly if anything thats providing the rarest of things: lawyer game content. It could probbably be trimmed down though but I cant quite figure out how.

I mean, if someone is showing a history of intentionally faking records so silicons can't release prisoners, we can handle that on a case-by-case basis. The only actual concern I have with that would be if someone's doing it for the express, explicit reason to prevent silicons from ever releasing someone, but they would have to be doing it to every prisoner for me to care.

Doing it a couple times to someone you don't like or is a jerk is an IC issue that lawyers can help suss out. Doing it all the time to powergame is lame as fuck.

[TheBibleMelts]

Nonviolent prisoners cannot be assumed harmful and violent prisoners cannot be assumed non-harmful. Releasing a harmful criminal is a harmful act.

has been changed to

Nonviolent prisoners cannot be assumed harmful. Violent prisoners cannot be subsequently assumed non-harmful. Knowingly or in ignorance of clear evidence releasing a harmful prisoner is a harmful act. Silicons can use the crime listed in a prisoner's security record as basis to determine if a prisoner is violent or not - if the crime is inherently violent (assault, murder, etc), then the prisoner can be assumed to be violent.

in silicon policy