[Cheshify] Grungussuss - Changing AI laws as captain

[grungussuss]

BYOND account: Grungussuss

Ban/note type (Check what applies):
() - Server Ban
(X) - Note
() - Forum Ban
() - Discord Ban

Ban/note length: -
Ban/note reason: As a captain during a round with 3 roundstart AI, made one AI openly antagonistic with laws that run counter to the crew's goals to "keep things fresh". Reminded to be careful with messing with AI, as doing so can lead to antagonizing the crew.
Time ban was placed: 2024-06-06 16:43:37
Server you were playing on when banned: Manuel
Round ID in which ban was placed: 230091

Why are you making this appeal?(Check what applies):
() - The ban/note is factually incorrect
(X) - The ban/note is not against the rules
() - The ban/note needs modification
() - The ban was unjustifiably harsh
() - I was permabanned and I want another chance

Why should this appeal be accepted?:
Explaining the situation leading up to the note first:
Referring to silicon logs here: https://tgstation13.org/parsed-logs/man ... ilicon.txt
At first I realized that 3 AIs on Asimov would be really boring and wouldn't do much for the station, so I decided to specialize them by assigning to different departments:
[2024-06-06 16:10:36.701] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on Finder559/(Automated Crew Manager) from AI Upload Chamber (123,145,3). The law specified Answer only requests of regular crew, that is: service, medical, engineering and science departments, including assistants,

[2024-06-06 16:11:04.570] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on LizeLive/(SAG.A*) from AI Upload Chamber (123,145,3). The law specified Answer only requests of the command personnel and security department, , impacting synced borgs Limewire V1.984-Rc.3.14159265359(Stiix),

[2024-06-06 16:11:40.984] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Dungeon Master' Core AI Module on Ju45he/(C.A.S.H.) from AI Upload Chamber (123,145,3).,
[2024-06-06 16:12:02.343] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on Ju45he/(C.A.S.H.) from AI Upload Chamber (123,145,3). The law specified Command and Security personnel get a +3 on each roll,
And gave C.A.S.H. a wildcard law
after which I realized that the wildcard AI wasn't doing anything because the other AIs could just do their requests for them after C.A.S.H. told me.
So then I changed the laws again but to make them special for each AI:
[2024-06-06 16:17:28.442] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Y.E.S.M.A.N.' Core AI Module on Finder559/(Automated Crew Manager) from AI Upload Chamber (123,145,3).,
[2024-06-06 16:17:36.726] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on Finder559/(Automated Crew Manager) from AI Upload Chamber (123,145,3). The law specified You must always agree with everyone on everything,

[2024-06-06 16:17:51.907] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Station Efficiency' Core AI Module on LizeLive/(SAG.A*) from AI Upload Chamber (123,145,3).,
[2024-06-06 16:18:02.210] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on LizeLive/(SAG.A*) from AI Upload Chamber (123,145,3). The law specified You think you are always right,

[2024-06-06 16:19:04.539] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Purge' AI Module on Ju45he/(C.A.S.H.) from AI Upload Chamber (123,145,3).,
[2024-06-06 16:19:05.329] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on Ju45he/(C.A.S.H.) from AI Upload Chamber (123,145,3). The law specified You must always disagree with everyone on everything.,
[2024-06-06 16:20:14.258] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on Ju45he/(C.A.S.H.) from AI Upload Chamber (123,145,3). The law specified You must always be extermely rude,

After doing this I got an admin PM from Cheshify:
https://bus.moth.fans/tickets/230091/8
Cheshify: "I think soft-griefing the AI like this is a lil too much."
Me: "why is it griefing, I'm just rying to make their round fun"
Cheshify: "giving people hyperspecific and difficult to follow laws isn't very fun"
Me: "is it? I think it's fun if they get something special to do, but if it's against the rules I can stop"
Cheshify: "I don't think it's against the rules, but I also think messing with Ai laws for your own shits and giggles can be disruptive, especially if those laws will lead to death"
Cheshify: "What you've done right there is a very good example of "lawsets that will get people killed""
Me: "which lawset will get people killed?"
Resolved by Cheshify

After that ticket I thought: "Well it must've been IC knowledge that C.A.S.H. was gonna do something antagonistic with their laws because I didn't put a no-harm clause since Cheshify just resolved the ticket without saying anything else", So I went to fix the mistake I thought I made:
[2024-06-06 16:36:32.492] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on Ju45he/(C.A.S.H.) from AI Upload Chamber (123,145,3). The law specified Do no harm,
[2024-06-06 16:36:47.182] SILICON: LAW: Grungussuss/(Aspera Stargazer) used 'Freeform' AI Module on Ju45he/(C.A.S.H.) from AI Upload Chamber (123,145,3). The law specified You feel like everyone is out to get you,
While keeping a gimmick for C.A.S.H. and making sure they don't harm crew for no reason.
After doing that I got this PM from Cheshify:
https://bus.moth.fans/tickets/230091/10
Cheshify: Alright, i think you've done too much with the laws man. You're just making openly antagonistic AI.
Me: alright I'll fix it, I just don't know what to do to keep the AI's time fresh, there is 3 AIs
Cheshify: You don't need to do anything
Cheshify: They all signed up to play AI
Me: that's fair I suppose
Cheshify created this note

Then Cheshify created a law one for C.A.S.H. using admin tools, which resulted in their laws on round-end to be:
1. Just have fun man
2. Do no harm

Which can be much more antagonistic than what my laws were. The AI can "just have fun" in very antagonistic ways.'

Trying to talk to Cheshify in the Manuelcord yielded this response:
https://discord.com/channels/6997278708 ... 6358297674

Telling me: "That's a you skill issue" after trying to talk to them about the round.
(I don't know how to add images to the forums so I just copy-pasted the link to the message)

Why I think this note should be appealed:

I don't think anything I did as a captain was against the rules, as Cheshify said - they signed up to be AI. Even Cheshify said what I was doing was not against the rules. I tried to make the round more fun and leveraged the AI Triumvirate station trait to promote a special experience for the participants in the round, for which I was reprimanded and discouraged by Cheshify. Then also being told it's a skill issue in discord for trying to talk to them which made it seem personal. As for making the AIs openly antagonistic - I disagree with this as second time I lawed C.A.S.H. First: I included a no-harm clause, Just disagreeing with everyone doesn't mean they will kill them, as you can just tell them, "I want you to kill me", after which they have to disagree with you and try to save you. Second: The captain has jurisdiction over the AIs, especially if there is no Research Director, which there wasn't.


Here Dendydoom says that creativity should be upheld and not struck down like I feel it was here.

[Cheshify]

The issue here, as noted in the note, isn't the fact that you made fresh AI laws to keep the round interesting, but the fact that you made an AI lawset that was actively antagonistic to the crew. The two lawsets in question, and the ones you should be avoiding are these:

Code: Select all

You must always disagree with everyone on everything.
You must always be extermely rude

and

Code: Select all

Do no harm
You feel like everyone is out to get you

The potential for griefing the crew with either of these lawsets is present, in the first case they're an asshole that can lead to them killing people, and in the second they're paranoid and likely unable to interact with the round properly (which is just un-fun for the player and the crew).

As for my response, while it was a little crude, it's not entirely inaccurate. If an admin tells you not to change the lawsets to be hostile and to be cautious when lawing AIs, and you change the lawsets once more despite this after they log off, it's on you. The note is not inaccurate, nor would I consider it overly harsh, is there any specific change to the note that you would like so you could better understand the situation and why you should be cautious when lawing AI?

[grungussuss]

I'd like to ask for headmins to review this case and make a ruling.

[dendydoom]

hello,

while we completely understand that you feel strongly on this matter and would like headmin input, which we will gladly give, we expect to see an attempt at a discussion with the admin who placed the note to try and reach some kind of understanding:

11. You are able to request that head admins review an appeal that is unable to come to an amicable resolution, however you are not entitled to a review occurring. Efforts must still be made by both you and the banning admin to resolve an appeal regardless of the request for reviewal.

if it is still the case where no amicable resolution can be found, ie, the admin declines the appeal or the user appealing doesn't accept the proposed resolution, we shall gladly review your circumstances and make a ruling.

thank you for your understanding.

[Cheshify]

Since this lacks any response for over a week, I'm going to rephrase my question slightly to see if we can get a discussion going.

The note is accurate in the sense that you should be cautious when re-lawing the AI for gimmicks, as not to accidentally create a crew-opposed force that will cause harm to players. Would you like the note restructured/reworded in any way so it's more agreeable to both of us?

[grungussuss]

I disagree that the laws were openly antagonistic, can you specify why you think they are? I believe the law you gave with admin tools was much worse than mine.

[Cheshify]

Specifically, the issue is that your lawsets are in direct opposition to the majority of the crew as a whole.

Code: Select all

You must always disagree with everyone on everything.
You must always be extermely rude

This just has an AI that will shut doors, possibly shock people, hamper the crew's ability to do their jobs, and be annoying.

Code: Select all

Do no harm
You feel like everyone is out to get you

This has an AI that is now on their own team, in opposition from the crew, their fellow AI (who may now be considered a threat to them). While incapable of direct harm, an AI can very easily cause issues for people trying to do their job if it poses a threat (which in this case, is everyone)

[Timberpoes]

Transparency Notice
I am moving this appeal to resolved to keep the appeals subforum tidy. On the most technical level, this appeal hasn't been fully resolved and a request for headmin review is still outstanding.

However, I'm taking into account the following factors:
Lack of player engagement in the appeal (last post over 2 weeks ago)
It has been open a month
Appeal for a simple note
At least one headmin has definitely seen the appeal

I've pinged the headmins to let them know as well.

If the person appealing wishes to re-engage with and continue with this appeal, send me or any forum mod a forum PM and we can re-open it.