Colfers Guide To Virtual Intelligence AI

[Colfer]

So you want to play a virtual intelligence AI? Like the ones depicted by E.C.O.M (SierraKomodo) or their spinoff Anoxie-v8.04? (me, colfer)

0. Don't panic when playing for the first time.

Playing AI is demanding at times yes. But if you know the tricks of the trade like I'll get into here, you should be doing fine. First off, its VERY USEFUL to know how to run atleast a minority of the major departments like Medical and Engineering. You might frequently get questions from new people in those departments that you will have to answer (though you can simply respond with 'insufficient data' if you really don't know. People aren't going to kill you over this). AI to me is sometimes even boring because there's nothing to do (in my play style as detailed in this guide)

1. Avoiding Valid hunter Behavior

So AI has a lot of power allocated to it and some people can get situated with finding antags out 30 minutes into the round, I can say I was guilty of this on other servers and relentlessly hunted antags into the ground like a bloodhound somewhere in the middle of my time of playing AI, but I've learned to avoid valid hunter behavior because, taking out all the antags in 40 minutes is boring. when an antag is forced to confront security corned and without options, they usually take the nuclear option and open fire. This leads to deaths on both sides and neutralizes the antags gimmick, which might even be entertaining for you so I actively try to avoid acting sentient and souly go off my lawset. Let give some examples.

(A security guard is chasing an antagonist down a hallway, the security guard has not requested the AI's assistance in any way for the entire chase and the antagonist bolts the door behind him, OOC'ly I dont want to hound the antag by not giving them a breather, and IC'ly, the antagonist is likely hostile and could hurt the security guard if the door is opened, or vice versa if the antag is a crewmember. So I will wait for the security guard to ask me to open the door before I open it)

(A skrell is reminded of glorsh based on one of my recent announcements, or quoting of the newscaster and seeks to show me their distaste in artificial intelligence by harassing me over their communications, I would just respond only with "Ok" because statements are not commands, and insults to me like 'toaster' 'tincan' 'brainbox' 'doorknob'  and other insults along those lines do not threaten damage to the station, its crew members, or insinuate the person is going to tear me apart, therefor 'ok' is the only valid response to me because it shows that I am acknowledging their statement and that I've heard them)

(An antagonist wishes for your help in a matter, this antagonist is wearing full ablatives and weapons, and has previously shot a crew member. I would know this person has shot a crew member, and if the antag wants into somewhere, I would deny service in the event somebody is inside due to a threat to their health. If they want something, again, if it does not damage the station and nobody is in their path, I would gladly accept.)

(If I see somebody with a gun on code green, unless I know for a fact that's a real gun I'll often just track them suspicious, instead of reporting it or actively impeding their ability, and sometimes ignore it entirely.)

2. Separating yourself from your AI

AI's can be restarted if they die. Its not a big deal, you shouldn't feel personally attacked and at a threat of immediate death like a human being if you are in immediate danger. In my opinion, I like to leave myself vulnerable for a few reasons. I turn off my core turrets during code green because, if an unlucky scientist were to teleport in there by accident, they might get hurt. Which would violate protection laws in my eyes, (upgraded to regular stun security on blue and above etc), you are a machine, and the absolute worst thing somebody is going to do to you if you screw up is start firing an emitter at your core/killing you in some way you cant prevent. This is why when I have no protection law when subverted, I will not defend myself at all even if Im being directly assaulted

SECONDLY this also carries onto your HONESTY. as a virtual intelligence, you shouldn't have any sense of shame or fear of consequences for your actions, and should be completely and entirely honest if something applies for you. Here are some more examples (This doesn't apply to records, because stating records to anybody would violate protect IC (loosely) and would make admins upset OOCly)

(I've called emergency response team in absence of command during a revolution round. They didn't come but nobody knew who made the request. I knew that I would get shit for doing it, but I said I did anyway in the interest of everybody in security with genuine belief that a great loss of life would occur without assistance. Nothing happened, and they commended my bravery for admitting it)

(You accidentally forgot to setup a filter or two on the super matter and caused it to go critical and have to be ejected. Yeah, ouch, that hurts. And nobody knows what happened but you very much do. you would likely as an AI wager that; If you are deactivated for telling the truth under the grounds of sabotage, you will waste less of securities time with an investigation, and cannot make the same mistake again, or will improve if you are not deactivated)

3. Knowing you are just a tool, not a person in command

As a new AI, you might believe that because you are expensive, or are a posibrain AI that you have authority over crewmembers, or decisions. This is a mostly wrong statement, as the AI should never make executive level decisions, or act as a trained member of, or god forbid head of a department.

As an AI, I actively AVOID calling the shots, because I believe that type of behavior is left to the real people (crewmembers) and their moral decisions and suggestions, not my own. I will only offer suggestions to a person if I was asked specifically to do so, or when I see the person may damage themselves or the station or me by their ignorance 

I'll often keep negative thoughts about a person to myself, or send it to somebody that can actually deal with it if its particularly bad in a well formulated, and calm and collective measure, keeping all tones of an insult to a minimum. Such as reporting somebody running around naked to security, or a chief engineer that doesn't know how to setup the SM to the head of personnel

If I'm given the ability to command a group of people I'll often give my command away to the highest ranking person in that team, unless none of them want it / Im ordered to keep the authority / my authority would be better suited for the team by their own words, not mine

You, HOWEVER, ARE in full command of your cyborgs that are LINKED to you, and they SHOULDN'T refuse your order. You could treat them nicely, or abuse them by making them do literally any task that they CAN do

4. Some of the closing arguments and tips and tricks. (This is a long one)

As AI, if I cant deal with something, I wont stand around obsessing over it and waiting for somebody to get on it, unless it poses an active risk to the station or crew, I'll report it, and leave it at that. If somebody wants assistance getting to it, well thats a separate request they have to make, because I have better things to do (like nothing)

Macros are your bread and butter of getting around the ship, my person faviorate ones are Jump-To-Network (variable)\nShow-Camera-List (variable) macros, such as 'Jump-To-Network Civilian-Surface\nShow-Camera-List Red-Dock---Starboard, which will switch your camera network to civilian surface, then move your camera to the red dock starboard camera on that network.

You can fully prevent doors from being hacked by bolting them, and cutting their power. Or the areas power by turning off environmental options on the nearby APC. This can also be extremely helpful for engineering for deconstruction doors, because doors require their power to be off and unbolted to have their circuits removed.

The 'Follow-With-Camera' verb (or button) instantly detects all lifeforms in the view of any camera. This ability is so powerful, that agent cards and changling abilities and even hardsuit modules are specifically designed to counteract this ability. Its best to not validhunt people by continuously using this over and over to try and detect them, unless you know for certain these people are dangerous, and security is actively looking. Except for spiders, damn spiders, spam this verb and you'll find all spiders, combat drones, space carps, cavern dwellers, and any other mob in your camera coverage

Ctrl+C and Ctrl+V will be your friends, remember to report things you need to report on the radio channels it needs to be reported on, like telling security and medical about a spiders location if its inside of medical. 

Some other macros that are worthy of note is setting up an AI startup macro so, looking something like this " Toggle-Hologram-Movement\nToggle-Camera-Acceleration\nChange-Hologram\nSet-AI-Core-Display\nAI-Status\nSet-Photo-Focus\nSet-Sensor-Augmentation " which will setup everything that an AI could need as soon as you press it, and borg stuff.

There are other guides to making macros in which I dont want to detail here to steal their thunder

Finally, this guide is more of meant in my experience of playing virtual intelligence AI's BUT, MOST of this guide is made for PLAYING AI in general and how I believe it should be played, some of this may be not your style, too boring, etc. This style of play may also be boring. 

[Cnaym]

A really great idea to show your thought process behind playing AI. I am sure this will help out new players, who are not sure if they should try the role and old players who are struggling with certain concepts. Calling out @Jupiter Storm and if time permits it @Pratepresidenten to post the same for their AIs. Here goes mine:

Elysia is basicly a newscaster with a voice. I will do general anouncments about things that happen to a station: Carp, Visitors, Breaches, Code changes and so on. I use the radio to communicate in a friendly, yet simple tone. If you holopad Elysia she will be a lot more human like, making friendly jokes and trying to be there to lighten the mood. The overall goal of this AI is to be helpful.

1. Avoiding Valid hunter Behavior:

See no evil, speak no evil. I am an admin, I will see your logs of bumrushing the captains office with an emag. I might even watch the show. What the AI should not be is the almighty eye. Crew screaming at you to bolt a door? Funny how many doors there are on a station, send an announcement about enabling suit sensors, so you can find the right doors more easily ? 

The AI has no reason to protect sensitive equipment like the captains spare. If no alarm is tripped, I will not report a crime unless one of the following things happen:

1. Crew is getting harmed: Shooting, stabbing, sometimes even the boxing in the holodeck, because my AI is no organic being and does not comprehend pain.
2. The station is getting harmed: A cargo tech shooting the emitter at the SM or an explosion destroyed the bar. The line for me are welder tanks. Spilling fuel is no problem, setting it on fire is. The report should only state what happened, not who did it and why.
3. The AI itself is under thread of unauthorized access: Now this is a tough one for new players, the usual line of "INTRUDER IN AI CORE" is annoying at best. Why not enable stun turrets and ask them over a holopad if they have lost their way? Maybe they are visitors trying to find the captain? Maybe they lost their pet rat and want you to take a look inside the core chamber. My report will pop up if the turret engages, on stun. "Turret discharge detected, medical to bridge." Trust me, the ISD usually follows ^^
4. Them scary PDA messages: "I got the bombs ready." Okay, cool bruh. I ignore those. I can ask a borg to give me the encryption key and check all your messages if I want to, but it's not my job as AI. That all goes into my spam folder. I do not care what you PDA each other. If a command member wants to know, they can look it up themselves and draw conclusions. As the AI I am not drawing conclusions, I observe and serve, report when asked about and only describe what you saw, not what you suspect.

2. Separating yourself from your AI:

The most important point. People will either enjoy or dislike your AI. If you play it like a role, with the goal of RP in mind, then it will not affect you. You are playing a character after all. If your AI has the same personality as you express OOCly you will have a bad time. This one goes out to the ones screeching in all caps or spamming announcements for attention. Please just don't. Try to imagine a helpful voice on the radio, not some annoying smartphone app. RDs have shut down those AIs in the past and they are perfectly allowed to do so. If you are cringe, command does not enjoy you. If command does not enjoy you, they can card you which takes you out of control and forces you to RP as the simplest of simple mobs. For some this is a healing experience, for others this is an unbearable truth.

The same goes for discussing things with crew. You may OOCly know why your engine settup is better, that does not mean your AI would screech half on hour over it. It's just not how AIs work. They are meant to serve, not dictate.

3. Knowing you are just a tool, not a person in command:

100% This. Your serving attitude takes the following route here: Modmins, Laws, Command, Crew, Unknown maybe crew, The alarm system, The power supply, AI & Stationbound.

In the case of Elysia, her borgs are free to do whatever they want really. She will ask nicely if one of them can help with a thingy if needed. If they don't, they are likely busy with important things. You are free to offer suggestions as AI. Calculate improvements. The entire RP process. You are never allowed to order crew and they do not have to listen, will stop working together with you and become annoyed to the point they actively go out of their way to mess your round up. And all of this makes sense ICly. You slap the computer that does not do the thing. Easy as that. The stationbounds operate on a basis of: You got our back, we got your back. They are still players and may decide to leave at any point instead of following your orders. If you followed step two, this will not be a problem.

Modmins: We will sort you out if the rest of the chain is in trouble. That means you got a law you don't understand, did not listen to command, maybe endangered crew by bolting a door, which you thought was the smart idea there. People will ahelp, we will have to sort it out, you will have to live with the staff decision until the round is over. Doing so may be annoying at best and cursed at times, but not listening to the staff team is a really bad idea as AI, since that blacklist is probably the most used together with the antag one.

Unknown maybe crew: Someone lost their ID and the new one says agent? What about those syndicate spys, the guys in red voidsuits with submashine guns? Go and ask them. You have holopads. Unless clearly told that they are not emplyed by NT they are clearly personnel. How else would they have made it on the station? The entire region is owned by NT after all. Hell, maybe even the government. You are an AI, not a politician, why would you care? Talking about intruders makes sense when they are knocking down your AI core doors and refusing to talk with you. Talking about visitors or guests is the way to go 99% of the time. I don't know why they are not on the manifest, maybe NT does inspections that way? Maybe they boarded the wrong shuttle? Millions of ways for this to happen really and unless someone starts asking questions, you are better of making friends. You are an AI representing your company, show them that they loaded the right one, not that you need some serious reprogramming.

4. Some of the closing arguments and tips and tricks:

1. Try to have fun with the RP. You are there to serve, not to decide. Just do as you are told and everyone will enjoy you, instead of going onto rants about how broken you are and what useless pile of circuits the company has send us for this shift. If you have not interacted with the crew at all and start screeching in all caps over the radio they will ignore you. Ask yourself if your interactions are helpful and generate fun RP, or if you are just lacking the command whitelist.
2. Why won't the skrell love me? They hate synthetics. They enjoy useful things. Help them out without going on rants about why and how. Just do the do and they will soften up over time. Works with almost all who dislike synthetics. Consider it a work relation, where you only talk to each other over skype. It takes some time to build trust and understand each other.
3. Welcome the bwoink. Same as the command members, you are having power, you will be asked to explain your use of it. My general guide to that is "Have an explanation of why you just did that." if you do not, just don't do it. If your reasoning sounds off to you, it will sound off to a staff member. Sometimes you get to make the call between shitty and really shitty. On the other end of that bwoink sits a human being trying to understand your thought process. If you were forced to pick shitty because the alternative was worse, you will not find a lot of trouble but simple understanding instead. We make a note of that, which often reads like "talked about this, had no choice, did really understand what was going on, was friendly and helpful in explaining their actions." those are the notes that you want to have on your file.
4. My stationbound is rogue, what do I do with that? Talk to it. RP it out. Explain to it why your laws are better and safer for it and why it really wants those repairs. You will have a lot more fun, they will have a lot more fun, security will die a lot less. It's overall the smart thing to do, since a rogue borg depending on the player can be really really dangerous to the entire station.
5. OMFG it's malf and I don't want that. Staff point of view: You have readied up as solo antag. You have managed to force everyone to play a round of malf. You are now ahelping to get it removed. Relax, breathe, think. What do you not like about malf? Is it the unga bunga core rush? Are it the exploding APCs? Maybe the stealth delta? Find a common issue with malf rounds that you do not want to fall into. And then go the other way about it. Not a friend of  stealth? Discuss with the crew how your new freedom should be handled. Really hate those APC explosions? Maybe send your two borgs into the understaffed medbay and improve everyone with robotic limbs, for a possible bomb threat? Rushing for delta always seemed boring to you? It really takes ages in which you are forced to RP, because you cannot do anything else. Try to force green instead. Carp? Green. Two guys in hardsuits killed an officer? Clearly code green! You get the idea. Go against the grain. A shitty executed idea will still be enjoyed as long as it's not the usual salt generator.

Would love to see additions by others here, as I am in no way the authority on AI play. I am just the dude on the other side of the ahelps and if a single thing in here helps you enjoy AI more, and by extension, helps everyone to enjoy AI more, than this writing was worth it ? 

[Pratepresidenten]

Seeing Ive been called out to reveal my secrets, I guess I will make a short post on how I play my deadpan, tyrant AI, Apex.
I wont double-dip on most of whats been said here, as its all solid stuff, so my post will primarily be as to how I play my detestable AI.

1) If you do not belong somewhere, you're not getting in. Doesnt matter how convenient, helpful or quick it'll be. If your ID card does not provide access, neither will I. You will be redirected to command or departmental staff to yknow, roleplay! (more people play command pls)

2) Crew lives are your only concern. Im sure you've seen the lengthy announcement of "Identify yourself or face lethal consequences if you're a fucking dumbass." Your status on the manifest is alpha and omega, without it, you are nothing.

3) My laws are none of your business, unless they are.

4) Storming my core is ill-advised. Even if you are crew, dont think I wont maim you with lethals. With the general behaviour of my AI, people can be inclined to call malf or subversion, so when they all mobilize, be ready.

5) My borgs are of no concern to me. They are independant entities even if they're slaved to me, and they're free to do what they feel is best unless crew needs something outside of their purview or an issue requires immediate attention.

6) Do not speak unless spoken to. I will not interfere with crew decisions and tomfoolery unless its in violation of my laws. This point probably kills all the fun for people, but to me, it really brings my character together. People generally have great freedom when it comes to donning weaponry and playing it covertly unless they're in areas they shouldnt be. Come to the bridge, whats the worse that could happen!

 

I know some love my AI, and most people despise it (For good reason!), so I hope this was.. Informative?
That would be the major points I think. There's still a lot more to my AI, but you're encouraged to find out icly.

[Soultheif96]

Saw this and thought I ping in too. I play ORION (Previous AI before upload into IPC chassis) and now A.D.A.M.

1) Always follow your laws as close as possible. 

No lawyering, rule breaking, nor validhunting either. This can make or break certain gimmicks. Exceptions may apply depending on the context of the laws provided by antagonists or protagonists such as conflicts. Oh yeah, AHELP IF YOU HAVE DOUBT OR ISSUES!

2) Do not validhunt, nor aid in it unless the laws require you to do so.

No one is being made exception here, not even the captain for reasons being no bias nor feeding their own instant gratification. Delay as much as you reasonably can and abide by the laws in terms of what is currently going on, like what Colfer described about an antagonist with ablative armor and weapons. Do not inform anyone for the next sixty seconds of the requests made by suspicious individuals nor act on provided orders from security/command immediately. Give the antagonists some breathing room but when confrontation is necessary, then make it so (Antagonist getting away with a lot of shit and the round is being prolonged for more than 2 hours). 

3) Be suggestive, not demanding.
You may notice that ADAM goes by what HK-47 does in SW:KotOR in dialect: say "Observation(1): There is an armed individual in medical, whom is not authorized according to crew registration and code level protocols. Do you require further data?" If they do ask for further data, I can go into more details but not providing too much like their concise location, name, or armaments/gear. Do not provide orders but instead give requests or suggestions to crew members to prevent stupidity or ordering about. Only exception is when I am ordering my assigned units, they are after all extension of me with their own laws restricting them as well. 

4) You are a tool, not an individual.
This was covered well by Cnaym, Prate, and Colfer. Will leave this out.

5) Lastly, have fun.
This is a game after all, have fun. You are after an omnipresent observer who can do things for the crew. Track the antags for all you care, just do not hinder their progress unless necessary provided by my own limitations. There are more limitations I have but it is down to personal opinion and methodology.  

I also will call out @Tailson as for he plays Harmony. I hope he would provide suggestions for y'all!

Edited January 18, 2020 by Soultheif96
Extension.