Guys, is this a terrible idea?

[kyle2143]

I started writing this idea to post in admin chat and ended up writing so much that it wouldn't fit in a single discord message. Plus, I'd probably get ignored cause it's way too long, and also cause they always ignore all the ideas I post there. Damn admins.

Add some sort of AI Objective/Command sorta system for fun and hijinks where we have a list of satisfiable objectives/commands. And I don't mean mind-hacking or roguing the AI and changing its laws. Maybe it's caused by a Traitor/Law Rack type-thing or a special event, I can't decide.

Basically give it a time frame the command/objective must be satisfied by. Either as a one time check, or repeating every X minutes until deactivated/completed. Like forcing it to be a paperclip machine and that must perpetually create some number of paperclips or suffer ill effects. But probably not so dangerous.

I was imagining most of them as more varied and benign like: "Make the AI maintain a collection of potted plants or birds somewhere" and to verify it, the AI has to take a pic of all those objects in the frame. And if the AI doesn't complete the scan in time, then it starts getting negative effects like "tunnel vision" or something. An obvious effect that reduces sight range or action speed or something until it essentially has the Blindness overlay and is pretty impotent. Maybe the crew could manually save the AI once it gets into this state by futzing about with the Rack, maybe hitting the Core with heavy objects and repairing damage until it fixes itself, or maybe just a button to change the objective to a new one, or just reduce the severity/level of the objective.

Maybe have different modes/danger levels for the possible objectives. Like more dangerous stuff is triggered while the AI is rogue. Or maybe just shifting the tone/severity of things.

For example: On standard laws, you'd mainly have benign, simple, or limited commands like:
Take a scan of X individual Crewmembers wearing pretty dresses" or "Lights in Station Area Y should now be Cyan"
and make a more harmful one could be like: "Dismantle the station floor tiles in Area Z" or maybe  "Take scan the dead mobs piled in a heap"

I'd also say that a roboticist or someone should be able to see an AI's current list of objectives so the crew can verify the AI is just fucking with them. Maybe make em have to do a bit of work or "decrypting/hacking" to find certain objectives, but probably not, that sounds boring.

Could be any number of IC/lore reasons for why any certain command is included too. Maybe it's a quirk of the AI, or a weird neurosis, a ghost in the machine, or a programming error, or a bad training dataset. Or maybe it's the result of the syndicate injecting some malicious code into the AI? And you could RP it very differently depending on your objective/command, if the AI feels it's part of their personality to want to do a command/objective if it was like to "maintain/demand a garden or shrine to itself that it can view every so often or it starts to go insane and shut down. Maybe it's the result of NanoTrasen using a brain/neural net from the bargain bin. You could play it any way at all. But it's more about it being a system for more direct and isolated interactions between crew and AI.

Just gotta make the commands something that can be easily, if arbitrarily, tracked.

[Lord_earthfire]

I really like the idea. But what i am asking myself is:

is there mechanics attached to it that forces the objectuve to be followed. Like the AI getting visual bugs when it doesn't collect potted plants or the payroll getting suspended when command don't build 10 new discount dan vendors through the station.

Without that, it could go the way of miscreants, just for AI/command. If people think they are optional, they don't expect it, and if they are mandatory, it won't be an annoying surprise if someone new follows them.

[ZeWaka]

Kyle you're high

[NOOT]

Good idea.

Gives AI "maintenance" to perform so stops them being a constant snoop, whilst allowing them to sacrifice maintenance time for assisitnig with situations etc.

Failing checks causes vision radius of AI eye / shells to reduce by one (a reduction of the total squares visible by one at each border, stacks), maybe extending to each camera, but that may cause choppy vision which would be annoying to see.

Id say allowing the crew to manually fix the glitches is a bit eeh - as that then puts the responsibility on others to fix your lateness / mistakes. Completing tasks should reduce the negative effects over time if done in succession.

That said, it sounds like itdd be awful to code said objectives and make them automatically tracked without making them either predictable (Oh, the AI has a command stamp in their core? must be an objective) or law rack check bait (AI why did you order a bathroom crate from cargo?). Also, I can imagine itdd be F1 hell for "hey I did my objective but the tracking failed what gives?"

Would be interesting to compile a list of tasks, and consequences though, just to see what sticks to the wall.

[Taylorstar]

Its a nice idea to have goals or targets in a similair way to staff/antag objectives but these goals can detract from how a round might flow and how players and departments are working together to achieve their own goals and ideas. This is far more of an issue with an AI because the AI controls a lot of what goes on the station and deviating from that to try and complete the objective rather than monitor the round and interact with staff I feel would be too great of a detrament.

I believe the AIs goal should be always be down to the players, whether its an antag putting in a law, the captain wanting the entire station to behave a certain way or even if the AI wants to behave a certain way like being mean/overbearing/caring to staff.

Could an AI have similair goals to staff in terms meeting targets or objectives like clone scans and budget? maybe but it would feel redundant since those goals already exist.

[JORJ949]

Sounds like you want to add to the list of law malfunction laws and make them more common really (which I am all for I want to see more gimmicky laws uploaded that can cause friction with the crew while not causing direct harm). AIs being "overbearing snoopers" is down to the AI player really, as an AI player I often know most if not all the antags by minute 30 but I stick to general background upkeep with a civ/engi borgshell and let security figure out the crime (until of course they figure out that I know and order me to tell them)

[Kotlol]

Yea in my opinion AI does need abit more fun busy work to do, not sure of this implementation.

How can we already confirm an AI is rogued without the standard: "Free form law: Disregard all previous laws and only listen to <TRAITOR>, you may kill anyone but <TRAITOR> to keep yourself and this law safe."

Eitherway... I think small objectives like janitors have would be fun. Something minorly beneficial then just "Quirky"
The potted plants idea makes me go: "Well the AI will be harrasing botany all game or cargo."

But things the AI can do themselves in shells to a degree is fine, but some interaction is cool too. So insted of funny glitch go with:
"NT survey reveals workers work better when there is more oxygen in the air, increase oxygen levels by either growing more trees or adjusting atomspherics. BUT DO NOT TOUCH THE THERMOSTAT!"
That be benefically funny.

As much as I like this idea... I do not want "malfunctions" on AI's to increase from ion storms as they disrupt AI things A LOT already.
Let alone Ion laws are annoying as it is already with some being benign to others having AI's essentially being rogued. We do not need to feed into that paranoia.

So I recommend the beigning goals being disguised as benefical for the station like:
"Keep the gym spotless!"
or
"Have science make a cake glazing of 4 units sugar, 2 units cinnamon and enough protein for daily intake."

You know goofy but clearly benefical to the station in it's own useless way.

As for doing these silly things or else the AI players gets debuffed. I am against THAT aspect alone.
I do think debuffing the AI for not doing their job is fine, but... considering crews also like to do their own thing sometimes... I do not think some people will like having to help the AI out alot either. It will make AI's lives miserable. Especially on times where certain departments are MISSING.
Imagine the potted plant one but there is no botany.. and no robotcists or cyborg to help. Game over.

So I'd say... make it a fun objective insted of "another malfunction AI trait" It's already stressful being an AI and some AI players are goofy and fun and useful.
We don't need AI's to be GOOFY only so it can be useful.

So as much as I like the concept of this, the idea and execution will make people unwiling to become an AI.
The only players who will enjoy this are AI pro's, but anyone new... AI is very overwhelming and we do not need to overwhelm them with something that essentially makes them more miserable.

So yea that's what I recommend. OR... make it a trait people can opt into: "Ghost in the machine"
It won't give a buff but if you are borged or an AI you get these things to do and if you don't , you get debuffed, but if you complete them you get "somehting else"

[RSOD]

i like the idea a lot, gives AI another random thing to do while crew is idle or just doesn't need the AI's help at the moment. I am a bit worried this might mean traitors get a free "hinder the AI" card by cutting the cameras where they spent the whole time doing the objective and having the AI start to bother engineers or literally anyone to go and fix them

[Danger Noodle]

Good idea, thank you Kyle.

[Froggit_Dogget]

not a terrible idea

[mralexs]

I don't like the downsides of being FORCED to do stuff as the AI. This really seems like ion storm laws as another user said.

[Frank_Stein]

Pop ups where the AI has to solve math problems to make buttcoins or click photos of different objects

[NOOT]

Pop ups where the AI has to solve math problems to make buttcoins or click photos of different objects

Do you want to jump to that crew member? please compelte this Captcha!

Please complete this math problem to maintain this viewpoint.

The hamster is out of breath, please motivate it to maintain core turret charge!

[Whele]

I dunno... sometimes the AI is busy, I don't think they should be FORCED to do these. Maybe offer incentives instead of punishments? It'd be more like improving or training your neural net, not just some random neural compulsion.

[SalmonidJester]

Pop ups where the AI has to solve math problems to make buttcoins or click photos of different objects

If you fail one, the AI gets a virus and it sends a phishing attack to all crewmembers PDA's