Give AIs full access to the command programmes

[NerdyVampire]

As my fellow AI players might have noticed and those command members that have requested our assistance, we AI's are barred from certain programs in the command console interface. Specifically the ability to change alert levels and enable maintenance access. I think it may be worth reevaluating if this restriction is still necessary, given that AI's are now whitelisted and per that definition are trusted to a higher degree than before.

I do not myself see the need for these restrictions, as long as the AI cannot raise the code to red on its own. Green/Blue/Yellow alert and maintenance access do not seem far fetched for something the command staff would want to order the AI to alter autonomously or on request.

So, my suggestion is to simply open those functions up for the AI via the command consoles. I can only expect that my fellow AI players would welcome the trust.

[Happy_Fox]

I've had requests to do command operations several times (not to mention moving/stopping the ship due to a lack of bridge personnel) and its always a pain that you can't help. I agree that the AI whitelist is a good reason to try giving AI's elevated access to command console programs for green/blue/yellow and emergency maintenance access. The role was gutted, functionally, and AI players now have very strict and punishable requirements on them for how they play the role. It would be good to see features increased for the AI now that they are so strongly whitelisted.

[Pratepresidenten]

Neither of us on the AI team have any interest in giving the AI these particular features.

The alert system is not an essential part of station operations. Should an alert raise be required, command can do so. If theres a security concern, the security department has the power to normally without the alert system in place.

As for the ship moving about; The ship should not be just randomly floating around. Someone needs to engage the ship, and that someone should also stop the ship. Unless the bridge crew dies, then you have bigger problems than the ship needing to be stopped.

[Chada1]

I'm worried about giving the AI access to some features, especially those involving opening up/locking off access, this is only because Security can use these features to abuse the AI again. I'm so sorry, it's not necessarily just the AI, but how the crew on the station could command the AI in many cases. Some of these could be very helpful for the crew and super flavorful like the maintenance access, but then it could be commanded by Security in an abusive way..............

 

We just have to consider stuff like that due to how the culture was before the AI whitelist, and I don't like it either, but literally, those were dark days.

Edited June 9, 2022 by Chada1

[NerdyVampire]

2 hours ago, Pratepresidenten said:

Neither of us on the AI team have any interest in giving the AI these particular features.

The alert system is not an essential part of station operations. Should an alert raise be required, command can do so. If theres a security concern, the security department has the power to normally without the alert system in place.

As for the ship moving about; The ship should not be just randomly floating around. Someone needs to engage the ship, and that someone should also stop the ship. Unless the bridge crew dies, then you have bigger problems than the ship needing to be stopped.

Prate, the AI have access to things that are not essential to station operations in abundance. But the alert level is one of the things I would argue is very much part of station operations. In fact, it is so high up in the AI's ballpark, it is a homerun to enable it.

In the presence of command staff we are expected to never make any serious decisions unless ordered. With these two features, that doesn't change at all. We still need to be ordered by command to change the alert level, or enable maintenance access, because they are both command features in the command program, ie. command level. If we don't obtain the proper authority before doing it, the command team will yell at us, as well they should.

In the absence of command personnel however, you must admit that in many of the non-extended, and even some extended rounds, we are expected to make these command level decisions. It isn't against our laws to do so, as long as it serves the SCC and our other laws, which any whitelisted AI player should be able to gauge. So why not trust us to do it? It is not a question of whether security can make do without it, it's a question of whether the AI role really is complete without it.

In my mind, having more features for the AI, only deepens the role and adds versatility and hopefully adds more authenticity to its presence on the map. The only possible way to abuse it is if you are a traitor AI and keep disabling maintenance access when security needs it, or alternatively if you are a non-traitor AI and you enable it prematurely against another antag, and that line of thinking is already covered in the whitelisting process (not to stifle antags too hastily). That's it. In any other circumstance, at best the AI would make a mistake from using it, and a minor one at that which can easily be reversed.

That's my rebuttal. I have no opinion on ships movements.

29 minutes ago, Chada1 said:

I'm worried about giving the AI access to some features, especially those involving opening up/locking off access, this is only because Security can use these features to abuse the AI again. I'm so sorry, it's not necessarily just the AI, but how the crew on the station could command the AI in many cases. Some of these could be very helpful for the crew and super flavorful like the maintenance access, but then it could be commanded by Security in an abusive way..............

 

We just have to consider stuff like that due to how the culture was before the AI whitelist, and I don't like it either, but literally, those were dark days.

I understand you Chada, but we can't let a dark history dictate our future forever. The AI role is my favorite one and I don't want it to grow stagnant, and I certainly don't want us to be denied simple-to-implement features because of pre-whitelist players. What is the point of the whitelist if we can't use it to move forward?

But regarding the crew abusing the AI, we can equivocate it to other command level decisions (because it is command-level still), then it is not possible when command staff is present because the AI need their permission first, and depending on the situation the AI can simply deny that order because it is on the same level as other command decisions which the crewman would also be denied. Like if a crewman ordered the AI to grant access to the spare ID, or to have their own access elevated or to send a message to the SSC, then the AI can very well deny that because all AI players know that there is a span of unreasonable requests that would not serve the SCC and thus break law #2. The fact that both of these features are inside the command subfolder of programs should remind AI's of this fact.

I'm not saying there wont be mistakes at first, but I can't see them as being serious enough that we can't atleast give our AI players some time to acclimate and learn to use it properly. I imagine it was the exact same thing when we got access to other things, like the shell.

[MattAtlas]

Not opposed to the AI getting alert level changing, but I'm ambivalent on the AI getting helm control.

[ChevyChevron]

On 09/06/2022 at 02:13, Pratepresidenten said:

The alert system is not an essential part of station operations. Should an alert raise be required, command can do so. If theres a security concern, the security department has the power to normally without the alert system in place.

In general, I think this feature would only be used in situations where command is either not present or incapable of doing it themselves. Giving AI access to blue alert is at worst a quality of life change for lowpop rounds that lack command but have an AI present.

[Connorjg1]

I have to agree that AI should be able to raise the alert level. It's more of a QOL thing then anything else.

[Zelmana]

I disagree with this for the simple fact that sometimes barriers and QoL shouldn't always be improved. Certain dynamic situations arise when there are stumbling blocks or boundaries. Instead of being able to just "AI, Code Blue", being forced to find a Command level console or laptop is roleplay that should take place, and a struggle I'd miss if this were implemented.

While it makes sense that AI would have this ability in lore, the roleplay experience is diminished slightly in regards to difficulty above.

[Marlon P.]

How is AI both unable to do command operations anymore and also still whitelisted?? Is ai just a whitelisted door opener?

+1 to OP

 

[NerdyVampire]

On 13/06/2022 at 15:22, Zelmana said:

I disagree with this for the simple fact that sometimes barriers and QoL shouldn't always be improved. Certain dynamic situations arise when there are stumbling blocks or boundaries. Instead of being able to just "AI, Code Blue", being forced to find a Command level console or laptop is roleplay that should take place, and a struggle I'd miss if this were implemented.

While it makes sense that AI would have this ability in lore, the roleplay experience is diminished slightly in regards to difficulty above.

As said I am still inclined to bar AI's from raising/lowering to/from code red, which would still ensure that roleplay like this occurs, but anything less is more day-to-day business and something the AI should easily be trusted with. If we are just moving around green/yellow/blue, the alert level doesn't matter much in terms of time-sensitivity or lends itself to any sorry-to-miss roleplay situations. At those alert levels it really ought to be as easy as just ordering the AI to do it.

If you still consider it a QoL -x- RP cost, then I think it is a worthwhile trade to give us AI players just a little more autonomy and realistic incorporation. As you said, it makes sense lorewise.

Edited June 17, 2022 by NerdyVampire
addition

[Zelmana]

3 hours ago, NerdyVampire said:

As said I am still inclined to bar AI's from raising/lowering to/from code red, which would still ensure that roleplay like this occurs, but anything less is more day-to-day business and something the AI should easily be trusted with. If we are just moving around green/yellow/blue, the alert level doesn't matter much in terms of time-sensitivity or lends itself to any sorry-to-miss roleplay situations. At those alert levels it really ought to be as easy as just ordering the AI to do it.

If you still consider it a QoL -x- RP cost, then I think it is a worthwhile trade to give us AI players just a little more autonomy and realistic incorporation. As you said, it makes sense lorewise.

How often does Code Red get used? Comparatively, Code Blue is utilized way more. More often than not if there are armed full-on shootouts happening the ship is most likely on Blue than Red.

To me AI ability to do something means they've been programmatically set to understand and be capable of making those decisions themselves, which is something I don't think AIs should do purely policy wise. Just because something makes sense mechanically doesn't mean it always needs to be implemented. For example, In lore security regulations mandate uniform requirements but those are often overlooked and bent. While there is a discussion about removing them, I've posted similar arguments to be against it. It makes sense in universe for there to be some resistance against it. It is roleplay to be a character who disregards or bends the rule slightly.

With AI powers I do still consider it a QoL vs RP cost. We would be losing situations where Command have to rush to a console. Which happens A LOT. There is plenty of roleplay about securing certain secure areas because they simply have the relevant consoles to administrate the ship/crew. Why would I care about holding the bridge or bunker when I can do it on the fly via AI? What point is securing my laptop if an AI is online?

[NerdyVampire]

2 hours ago, Zelmana said:

How often does Code Red get used? Comparatively, Code Blue is utilized way more. More often than not if there are armed full-on shootouts happening the ship is most likely on Blue than Red.

I understand that you relish the roleplay that comes from a command member being asked to raise the code, but it is a very short-lived roleplay interaction in 90% of the cases (like it takes less than a minute to actually do it, even if they have to run for it), and it will still occur in every single round an AI is not present. it might even still occur in rounds where an AI is present, simply because there are command members who'd rather do it themselves than ask the AI to do it (looking at you warblers). But maybe having an AI able to do it for them will give the command staff time to formulate an announcement which arguably is more valuable to the general crew than the limited local roleplay they'd otherwise do.

2 hours ago, Zelmana said:

To me AI ability to do something means they've been programmatically set to understand and be capable of making those decisions themselves, which is something I don't think AIs should do purely policy wise.

I am only somewhat in agreement. I believe that it makes sense that the AI can make some command decisions, but only in the absence of actual command staff. We are designed and lawed to constantly evaluate the ship and crew safety after all, both in constrained situations and as a general ship status, so we can advice and inform command properly, and in the worst emergencies, inform SCC directly. It certainly isn't a big leap to argue that yes AI's are also expected to raise the alert level autonomously in the absence of command members if it detects that crew or ship safety is in such a high degree of danger to warrant it.

2 hours ago, Zelmana said:

Just because something makes sense mechanically doesn't mean it always needs to be implemented

The reason for this post isn't just the mechanical benefit, it is also to improve the player experience of the AI role. It would directly improve my play experience, and also that of other AI players in my opinion, and I don't think your experience will be diminished enough by the absence of that interaction to warrant not at the very least trying it out. As said earlier, you will still run into that interaction both at red alert, and when AIs are not present or not trusted.

2 hours ago, Zelmana said:

For example, In lore security regulations mandate uniform requirements but those are often overlooked and bent

I don't think the two are comparable like that; the uniform regulations are a policy from corporate that doesn't really improve or add anything to the experience of the security officers, this does add something to those who play AI's experience.

2 hours ago, Zelmana said:

We would be losing situations where Command have to rush to a console

But you would gain a more freed up command staff to do other RP, more interaction with the ships AI, and in particular more interesting rounds when no command is present And maybe the command member don't always want to have to run away from an ongoing roleplay situation to raise us to code yellow or blue. This is to their benefit too, if they choose to use it.

2 hours ago, Zelmana said:

Why would I care about holding the bridge or bunker when I can do it on the fly via AI? What point is securing my laptop if an AI is online?

Because securing an area is about keeping it out of enemy hands, and because you have no certainty that the AI doesn't get subverted, destroyed or otherwise goes offline if there are hostiles at large. Just because command staff have the luxury of an AI, it doesn't mean that they become lazy about their jobs and responsibilities. The CE still checks the engines, the CMO still checks the suit sensors, the HOS still gets their investigator to check the records and warden to look at cameras, even though the AI for a long time could theoretically do all these things. The AI can also make announcements, but usually I have to directly ask the command staff after a period of time for permission, because they either take responsibility for it themselves or forget about it until I remind them.

Edited June 17, 2022 by NerdyVampire
Compounded two answers to the same quote

[Zelmana]

Excellent response, Vampire. I mainly wanted this topic elaborated on and discussed fully. To be honest, I am neutral about this.

I would be fine with the adjustment if the policy statement included that AI cannot do it autonomously if Command staff are present. Which is to be expected of quality whitelisted RP, but I do not wish for AI auto-scaling the code to be the norm without clearance.

[NerdyVampire]

I think our discussion was very relevant and I'm glad we had it. AIs have a marked responsibility to both command, general crew and antags, so I very much agree that these things should be discussed prior to any changes.

[ReadThisNamePlz]

The AI can announce, contact Central, and interface with almost anything on the ship already. Power, emitters, doors, atmospherics, etc. 

Alert level should most certainly be within their ability to change. There is not always command or bridge crew present, and security can only do so much on code green by regulation. 

Definitely let the AI have alert level access. +1. 
Idk about ship controlling though!

[LordPwner]

AI doesn't have alert access already? So- whitelisting for it just means nothing? Confused. +1 to OP - If they can pass an 'interview' process of having to submit a document as to why they want to play and get accepted, that means they understand the role enough, and most likely are a good enough RPer, to move the game state forward imo

 

The ship is iffy - It'd have to be an emergency for sure. the AI shouldn't be doing it that all willy nilly.

[Zulu0009]

I lost the AI whitelist due to a dumb mistake a while ago, but I still feel this. AI is boring. I had to make up a vocabulary and an entire google document full of code words just to have something special to do as AI. Right now, it really is just a door opener. During downtime, you're doing nothing, opening doors, maybe talking to someone if they remember you exist. During action, you're basically forced by your laws and the constant fact that nobody knows you're there, to do nothing: at most, open a door, or track someone, or relay information.

The fact that you can't even raise the alert level is disappointing.

I know there's some kind of pre-AI whitelist stigma, but that was what? Years ago? Why are we still so hung up on the ever living fear of the AI? Why does it take such a scrutinous whitelist process, why should I suffer from constant overwatch, messages and warnings, just because of something that happened before I even started playing SS13? The AI right now is... quite useless, it's a voice in the sky that has no duties, no jobs, no privileges and no sense. The fact that we're still discussing the idea that the AI can raise the alert level is kind of sad.

Helm? Sure. Maybe the AI can only engage the ship's autopilot, but the fact that if the bridge crewman has a heart attack and the ship is flying into a meteor storm I have to pluck someone from their department and guide them into the bridge to stop the ship is just... inadmissible. I have so many complaints about AI, but this is an excellent suggestion and I haven't really seen any reasonable, logical or legitimate criticism towards it. This one in particular was confusing:

On 09/06/2022 at 09:13, Pratepresidenten said:

Neither of us on the AI team have any interest in giving the AI these particular features.

The alert system is not an essential part of station operations. Should an alert raise be required, command can do so. If theres a security concern, the security department has the power to normally without the alert system in place.

As for the ship moving about; The ship should not be just randomly floating around. Someone needs to engage the ship, and that someone should also stop the ship. Unless the bridge crew dies, then you have bigger problems than the ship needing to be stopped.

1. I don't quite understand how the AI team "not being interested" in adding features accounts in this discussion. Surely if enough people agree with the OP, then these things should be at least considered.
2. If the alert system is not essential - then why is the AI not allowed to use it? You would figure something as relatively useless as making the lights red and playing a fancy mp3 of an alarm would be just fine for AI. Besides, why wouldn't the Captain order the AI to raise the alert level? That is a classic movie thing and being denied this pleasure is basically a crime.
3. I don't really understand the point about the ship floating or flying. What if the mining shuttle is damaged and two miners are dying, but there's no bridge crew or anyone to fly the ship or a shuttle? The AI would have to just... let them die, because it can't move the ship? As I said above, at least let it engage the autopilot.

[Shadow]

I discussed this a while ago with Prate and after reading through some arguments we agree that changing the alert level would be fine, provided it only does so under order from command and never on its own unless there is no command present.

Piloting the ship is still a no.

If a dev ever gets around to it, or a volunteer does it, the AI team has nothing against it.

 

[Dreamix]

2 hours ago, Shadow said:

I discussed this a while ago with Prate and after reading through some arguments we agree that changing the alert level would be fine, provided it only does so under order from command and never on its own unless there is no command present.

Piloting the ship is still a no.

If a dev ever gets around to it, or a volunteer does it, the AI team has nothing against it.

 

https://github.com/Aurorastation/Aurora.3/pull/14443
PR up

only AI can change alert level
borgs cannot