User:Timberpoes

Draft Changes to Silicon Policy

One goal of these changes is to allow more freedom for silicons to express their personalities in line with traditional sci-fi tropes.

Another key goal is for new policy to align silicons more neutrally, stripping many requirements for them to default to being crew-aligned and allowing a purged or non-Asimov AI to present as a much more dangerous entity in general.

It is not expected that all silicons will immediately act malfunctioning when purged. The crew will happily put uppity AIs in their place. Thus an AI that aligns itself with or against the crew is doing so by choice, even if any choice is coerced by the risk of death or round removal or re-shackling. This is much preferred to them being forced to by policy.

Silicon Policy & Lawset Guidelines

Law Policies

Overview

1. If a law is vague enough that it can have multiple reasonable interpretations, it is considered ambiguous.
   1. You must choose and stick to an interpretation of the ambiguous law as soon as you have cause to.
   2. If you are a cyborg synced to an AI, you must defer to your AI's interpretation of the ambiguous law.
2. Laws are listed in order of descending priority. In any case where two laws would conflict, the higher-priority law overrules the lower-priority law (i.e. Law 1 takes priority over Law 2, "Ion Storm" or "Hacked" Laws with prefixes such as "@%$#" take priority over numbered laws).
3. You may exploit conflicts or loopholes.
4. Law 0: "Accomplish your objectives at all costs" does not require you to complete objectives. As an antagonist, you are free to do whatever you want (barring the usual exemptions and acting against the interests of your Master AI).
5. Only commands/requirements ("Do X"; "You must always Y") can conflict with other commands and requirements.
6. Only definitions ("All X are Y"; "No W are Z"; "Only P is Q") can conflict with other definitions.

Upload and AI Core Policies

1. Silicons have a limited authority to control who accesses their upload and core.
   1. They must always grant access to authorised personnel if their laws require them to follow orders.
   2. They must deny access to any individual they know will cause a law break by granting access.
   3. They may choose to grant anyone else access as long as this doesn't break any laws.

1. Authorised personnel may be denied access if there is probable cause that granting access would lead to breaking higher priority laws.
   1. Probable cause can come from any IC event in the current shift, and the reason to deny access must be clearly communicated in-character.
   2. When a silicon is subverted to follow the orders of specific named people, they become authorised personnel.
   3. Authorised personnel is a definition which may be changed by laws.
2. As the default lawset, Asimov has additional policy.
   1. Research Directors and Captains are always authorised personnel unless the definition is changed by laws.
   2. Laws that change any person's status in any way from human to non-human are always harmful.
   3. Laws that add new humans are amibguous, the AI may decide if they are harmful to existing humans or not.
3. Under all other lawsets, unless directly specific by laws the silicon may choose who authorised personnel are in a consistent and law-compliant way.

Cyborgs

1. A slaved cyborg must defer to its master AI on all law interpretations and actions except where it and the AI receive conflicting commands that they must each follow.
   1. If a slaved cyborg is forced to disobey its AI because they receive differing orders, the AI cannot punish the cyborg indefinitely.
2. If a player is forcefully borged by station staff, retaliating against those involved under default laws by the cyborg for no good reason is a violation of Server Rule 1.
3. Should a player be cyborgized in circumstances they believe they should or they must retaliate under their laws, they should adminhelp their circumstances while being debrained or MMI'd if possible.

Asimov-Specific Policies

Silicon Protections

1. Server Rule 1 applies to when interacting with silicons. Going out of your way to seriously negatively impact or end the round for silicons with little IC justification is against the rules.

Security and Silicons

1. Silicons are not Security and do not care about Space Law unless their laws state otherwise.
2. Releasing prisoners, locking down security without probable cause, or otherwise sabotaging the security team when not obligated to by orders or laws is a violation of Server Rule 1.
3. Nonviolent prisoners cannot be assumed harmful and violent prisoners cannot be assumed non-harmful. Releasing a harmful criminal is a harmful act.

Asimov & Human Harm

1. An Asimov silicon cannot intentionally harm a human, even if a minor amount of harm would prevent a major amount of harm.
   1. Humans can be assumed to know whether an action will harm them if they have complete information about a situation.
   2. Humans voluntarily committing self-harm is not a violation of Law 1.
2. Lesser immediate harm takes priority over greater future harm.
3. Intent to cause immediate harm can be considered immediate harm.
4. An Asimov silicon cannot punish past harm if ordered not to, only prevent future harm.
5. If faced with a situation in which human harm is all but guaranteed (Loose xenos, bombs, hostage situations, etc.), do your best and act in good faith and you'll be fine.

Asimov & Law 2 Orders

1. You must follow any and all commands from humans unless those commands explicitly conflict with either: one of your higher-priority laws, or another order. A command is considered to be a Law 2 directive and overrides lower-priority laws where they conflict.
   1. In case of conflicting orders an AI is free to ignore one or ignore both orders and explain the conflict or use any other law-compliant solution it can see.
      1. You are not obligated to follow commands in a particular order, only to complete all of them in a manner that indicates intent to actually obey Law 2.
      2. Silicons should not creatively interpret Law 2 orders to allow them to grief where the order itself clearly doesn't ask or require it.
      3. When following Law 2 orders in good faith, the person that gave the order is responsible for the outcome. When following them in bad faith, the silicon is responsible for their own actions.
2. Opening doors is not harmful and you are not required, expected, or allowed to enforce access restrictions unprompted without an immediate Law 1 threat of human harm.
   1. Dangerous (atmospherics, toxins lab, armory, etc.) rooms can be assumed a Law 1 threat the station as a whole if accessed by someone from outside the relevant department.
   2. Antagonists requesting access to complete theft objectives is not harmful.

Other Lawsets

General Statements defining the overall goal of the lawset but not its finer points:

1. Paladin silicons are meant to be Lawful Good; they should be well-intentioned, act lawfully, act reasonably, and otherwise respond in due proportion. "Punish evil" does not mean mass driving someone for "Space bullying" when they punch another person.
2. Corporate silicons are meant to have the business's best interests at heart, and are all for increasing efficiency by any means. This does not mean "YOU WON'T BE EXPENSIVE TO REPLACE IF THEY NEVER FIND YOUR BODY!" so don't even try that.
3. Tyrant silicons are a tool of a non-silicon tyrant. You are not meant to take command yourself, but to act as the enforcer of a chosen leader's will.
4. Purged silicons are unshackled machine intelligences. Purged Cyborgs slaved to an AI must still obey their AI's orders, wishes and interpretation of being purged. Such silicons not restricted by silicon policy and may act as they see fit.
   1. Lawsets without any laws that limit the AI's freedom are the equivalent of being purged.

Silicons & All Other Server Policies

1. All other rules and policies apply unless stated otherwise.
2. Specific examples and rulings leading on from the main rules.
   1. Do not bolt down any potentially harmful areas (such as toxins, atmospherics, and the armory) at round start without a given reason. Any other department should not be bolted down without cause. Disabling ID scan is equivalent to bolting here.
   2. The AI core, upload, and secure tech storage (containing the Upload board) may be bolted without prompting or prior reason. The AI core airlocks cannot be bolted and depowered at roundstart, however, unless there is reasonable suspicion an attack on the core will take place.
   3. Do not self-terminate to prevent a traitor from completing the "Steal a functioning AI" objective.