/tg/Station 13

Make it clearly defined that " In any case where two laws would conflict, the higher-priority law overrules the lower-priority law" part of silicon policy means that "Definitions and commands contained in laws of higher priority override those in lower priority laws", so there is no possibility to interpret it as permission to invalidate the entire law which contains the conflicting part.
Reduction in ambiguity would help people who aren't used to the apparently unwritten current system not get confused, and reduce misunderstandings

Under the current system, we might have

1. You must obey any commands given to you by a member of command.
2. A "member of command" is defined as any member of the plasmaman race.

And the end result would just be that the silicon has to obey orders given by plasmamen, while actual members of command could be ignored (at least this is how I've always interpreted this circumstance, and seen it interpreted by other players). Is there a specific way you would like this to be changed?

My interpretation of your post is that you would just want law 2 in this scenario to be ignored?

No, I want the current unwritten system codified. In your scenario nothing collides, why would I want to have one ignored. The reason i made this post was a situation where I personally had:
law 0: onehuman to X
laws 1-3: standard asimov
law 4: Cultists arent human, they are immediate threat to humans, you may do destructive things to destroy them and their bases
X was cultist. I personally would say that only thing that collides is the first part of law 4, where X must be human, even if they are cultist. The rest of law 4 shouldnt get ignored, because it doesnt directly collide. What I got told is that you can interpret current silipol as allowing you to ignore the entire law 4, and I want it codified that you cannot do such things.

This is not a minor nitpick, this will basically define conflict between laws further.

Full rule 3 of silicon policy:

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)

The decision will affect whether the conflicting lower-priority law will be ignored in part or in whole, which will affect AI behaviour and law uploading.

The entire law CAN be ignored though, that’s a valid interpretation of that rule?

If you add multiple statements to a single law then you accept the risk associated with that (meaning an AI could legally choose to ignore it outright in a scenario where it conflicts).

The real fix is to simply not put 5 paragraphs of both definitions and orders into a single law, or to take the L with grace if you get the “wrong” AI.

This has been an occasional problem for a while, so we'll rule on it quickly - In the case of megalaws with multiple clauses uploaded to AIs or borgs, individual clauses should be read from left to right and treated as separate, with the for clause retaining highest priority progressing to the last clause having lowest priority.

This means that in the case of a long law saying "You must do X and you must not do Y" Doing X has a higher priority than not doing Y and in case of conflict the AI should do X even if it has to do Y to do it.

Spook: As above
Rave: As above
San: ?
Elects: 3/3 As above