Status and Future of the AI

[Arrow768]

Status and Future of the AI

Over the last years the AI has been a topic that regularly comes up for discussion and various suggestions have been made to change, remove or nerf it.
This is mostly due to the ease with which the AI can cause issues with various antagonists.
Often all that is needed to prematurely end the round of a antagonist is that the AI spots them in an inopportune moment and announces their actions to the general population.

This resulted in the creation of a player poll about 7 months ago to ask the player-base how they feel about a removal of the AI.
65% of the votes in that poll were in favor of keeping the AI.

In last year we have attempted to mitigate issues caused by the AI by nerfing some of the capabilities of the AI. (Such as bolting doors, ...)
Even with these changes implemented the issues with the AI remain.

We have also attempted to increase the administrative oversight of this role with limited success.

However despite all these issues, the AI is also a unique and interesting roles that many players enjoy (as seen in the various polls regarding the AI).

Therefore we plan to approach this situation differently by requiring a Whitelist to play the AI.
Due to the difference in requirements, this whitelist will not be based on the existing IPC or command whitelist.
Instead it will be governed by a different set of requirements that are intended to ensure the desired play-style.
In addition a number of mechanical and AI-law changes will be implemented to support the whitelist.

 

The Laws (and directives)

• To allow the AI some leeway when dealing with antagonists the laws will be changes as follows:
• Safeguard and ensure only authorized personnel gain access to areas of high security or importance to the station and its operation.
• Serve and assist NanoTrasen and assigned crew, with priority as according to their rank and role
• Safeguard sapient life on board
• You are a valuable asset. You must avoid tampering from unauthorized entities.

To support the safeguard law, a additional directive will be added that lists the relevant areas for law 1.

The directive that governs the access to the AI core will be simplified and brought in line with other existing directives:
"In order to access the AI upload or the AI core a captain level decision is required. All command staff must be informed of the decision. A roboticist may accompany the head/s of staff if their technical skills are required."

 

The mechanical changes

• The following mechanical changes will be implemented:
• Standardizing camera wires
• Remove the vault motion alarm (currently being debated)
• Add Soft-Bolting to all airlocks (a few seconds delay + warning message on bolt, instant unbolt)
• AI core rework with a hard exterior and a soft interior to make it easier for antags to subvert the AI
• Remove malf (the game mode)
• Re-add traitor Ais with a possibility of an admin-given fully upgraded malfunction role (pending ability reworks)

We also plan to evaluate if some of the implemented AI-nerfs are still necessary after the AI whitelist has been established for a while.

 

The whitelist process

Unlike the species or command whitelist, there will be a Discord interview after the basic requirements have been met.
During that discord interview the applicant will be confronted with several scenarios that may occur while playing a AI.
The goal of this discord interview is to prevent the applicant from searching the forum archives / other active applications for the previous scenarios used during other successful interviews.
The interview allows the whitelist team to present the player with specifically crafted scenarios similar to how they can occur in the game while requiring a response in a timely manner (as it is common in the game).
The response of the applicant can then be evaluated immediately and the scenario further developed based on their responses.
It also serves to add a time-element to the interview (that is also present during general game-play and therefore helps to elicit responses that are more in-line with their actual play-style).
In addition a trial-period for the applicant to play the AI is being discussed internally.

 

My Thanks goes to @Pratepresidenten and @Shadow for their work on that proposal after given the initial idea to work with.

 

F.A.Q.
Q: Are the laws / changes / whitelist procedures outlined in that topic fixed?
A: No, they can change based on the feedback provided. You are encouraged to point out any issues that you see and suggest alternatives and additions that you think might be useful.

Q: Will there be a lore-part to the whitelist / Will someone from the lore team be part of the AI whitelist team?
A: No, this whitelist is geared towards the OOC and IC behavior of the AI players which is enforced by the modmins.
A2: As there is no requirement that a AI is aware of the events in the grater universe, there is no point in adding a lore-component to the whitelist. AIs that violate the established lore can easily be dealt with through the established administrative channels.

Q: Why the whitelist?
A: Because the current approach of whack-a-mole is a effort in futility. The average AI shitter joins, gets banned and moves on to another server. The whitelist requires a effort from the player to get it and thereby prevents the low-effort shitters from playing the role. (As an example: out of ~40000 players only 400 hold the command whitelist)

Q: Isnt the whitelist too much effort?
A: The interview is certainly more effort for the whitelist team, however the team has expressed the willingness to perform the interview.
A2: The applicant is expected to read and understand the documentation for the AI, play a certain number of rounds as a borg and write a application. The time required (~20 minutes) for the interview is a minor obstacle at that time.
A3: Once we have gained more experience with the AI whitelist, the interview can be reevaluated. However it is always easier to lower requirements than it is to increase them.

Q: Will there be a exception for long time AI players?
A: No. It would defeat the point of the whitelist.

Q: Why not tie it to the command or IPC whitelist?
A: The goal of the whitelists are different and it would defeat the point of the whitelist by allowing a large number of players, who might not be aware of the AI-expectations, to play the role.

Q: Where can I apply for the whitelist?
A: It is not possible to apply for the whitelist at the moment. We will make another discord announcement once it is possible to apply.

Edited January 5, 2021 by Arrow768
Added "Where can I apply for the whitelist?" to F.A.Q.

[Kintsugi]

That’s all well and good, but the real question is “when are we removing wizard?”

[SatinsPristOTD]

I'm looking forward to this, actually. It keeps my favorite spot, and governs it better. So 10/10 change, Aurora. I look forward to the process.

I hope you ping the Discord server when the applications are open, least I forever be in the dark. I never check the forums without poking.

[Faye <3]

I think this is definitely a good change, though my only concern is that AI can be somewhat rare (at least in my experience! maybe i am just unlucky) and that the whitelist could worsen that. But I think it is probably necessary. 

Edited January 5, 2021 by Faye <3

[NerdyVampire]

Love it. I think this will be a great change.

[Winter404]

I'm really really excited to see how this works out.
I love the AI role and it's been disheartening seeing some people hate it or calling to "just remove it". Specially when, at least in my opinion, the biggest problem came from bad players, rather than the role itself. Also, love the mechanical changes (both the ones proposed here, and the ones so far). I hope this means the AI will get subverted more often.

[Arrow768]

25 minutes ago, SatinsPristOTD said:

I hope you ping the Discord server when the applications are open, least I forever be in the dark. I never check the forums without poking.

That will be the case.

 

[EmperorVoyd]

Does this ridiculous change hurt my ability to play my borg Glitch? 

[Arrow768]

6 minutes ago, EmperorVoyd said:

Does this ridiculous change hurt my ability to play my borg Glitch? 

No, the borgs wont be covered by the AI whitelist.

At this time we have not discussed if we want to extend un-nerfs (that we plan for the AI) to the borgs aswell or if we want to keep the capabilities of the borgs as is.

[SpaceBaa]

Unless we lose all rights when we step outside of the station, I propose updating Law 3 to read:

Safeguard sapient life.
Or:
Safeguard the crew.

I don't know whether non-crew are considered vital to save, but I'm also not sure if it is only crew members that the AI must protect, thinking of the merchant and random NanoTrasen Officers that turn up, so could also go for something like:
Safeguard authorised sapient life-forms.

[Alberyk]

1 hour ago, SpaceBaa said:

Unless we lose all rights when we step outside of the station, I propose updating Law 3 to read:

Borgs should not leave the station. This was done to stop them from just raiding away sites with zero effort.

[Chada1]

I love this and I've been the champion so far for increased AI/borg responsibility and even the AI nerfs --

My only concern here is if 'borgs are remaining unwhitelisted, that they should continue to be unable to bolt, it is my favorite role to play and it was ridiculously negative to towards the game for them to be able to bolt like they used to, even with the delay I think 'borgs would be a massive problem in that area

So maybe leave the bolting stuff to the AI and remove it wholesale from non Antag 'borgs? I dunno, but I'd prefer it that way.

[Chada1]

I also think the laws could use some adjustment as outline in the thread too:
 

* Safeguard NanoTrasen facilities and ensure only authorized personnel gain access to areas of high security or importance to the facility and its operation.

* Serve and assist NanoTrasen and assigned crew, with priority as according to their rank and role.

* Protect sapient life on board. 

* You are a valuable asset. You must avoid tampering from unauthorized entities and destruction. 

The change from Station to facility will make the lawset not require a massive change in terminology for the next big thing, assuming the ship we'll be on is a NanoTrasen one.

Two safeguards will be harder to memorize and confused when people say 'Law (Insert number) safeguard' in OOC and IC conversations.

The removal of the damage clause from the old laws will return 'borgs to what they used to be, where they were allowed to rush into hostage situations as written since their existence didn't really matter. I added that clause so that they'd have a pseudo-fear RP and avoid lethal danger. Crew can still treat them as expendable and send them on dangerous missions around the station, the 'borg would just be expected to be seriously careful while doing it.

Any number of other ways of fixing these problems work too but I'm seriously worried by them if we don't make changes.

[Scheveningen]

I'm going to avoid the "well what if this doesn't work" comments since this is a bigger step than has ever been taken to try to make the AI mesh more with gameplay. I've been of the opinion for some time that the AI really needed whitelisting because it had roughly equivalent power and responsibility to a head of staff. I don't mind the separate whitelisting, it makes sense.

[Chada1]

O also I love the 'Protect/Safeguard Sapient Life on board' because it effectively cuts the AI and 'borg validhunting out at the source -- If a 'borg/AI is going out of their way to endanger/shut down Antags they're actually in violation of their laws with that change -- I think that's a good thing.

[KingOfThePing]

Initially I was a bit reserved about the whitelist and interview part of these changes, I thought it would drive people away from a very enjoyable role, if it is played well. 

After some talking I have changed my mind and I think this is a good approach and good mechanical and non-mechanical changes. While I probably won't apply for the foreseeable future myself I think this will improve the state of AI significantly.

Edited January 6, 2021 by KingOfThePing
Phone autocorrect

[Jupiter Storm]

Very happy to see this change after so long, especially after all the fearmongering about its 'potential' removal altogether.

It just shows that together with community support and passion for the game so many positive changes can happen, just like other well-loved additions to the game.

No complaints about a whitelist either, even for those who've played it before. ?

[Skull132]

41 minutes ago, Jupiter Storm said:

Very happy to see this change after so long, especially after all the fearmongering about its 'potential' removal altogether.

I mean. Removal was a legitimate option that was discussed for a few solid months. I am not sure what about being open about such discussions counts as "Fearmongering".

[Jupiter Storm]

1 hour ago, Skull132 said:

I mean. Removal was a legitimate option that was discussed for a few solid months. I am not sure what about being open about such discussions counts as "Fearmongering".

Around the time when I stopped playing it was largely all hearsay and rumours, and accompanied by some unpleasant and biased hate from an assortment in the community.

I'm glad that in more recent months it was discussed openly by the staff and people were kept informed. I had no idea all of this went on while I was away, so I'm sorry if it sounded related to that at all. Keep doing your good work! ? 

[MEDTECH]

I like this change, it gives more meaning to the AIs, and save it from deletion.

But I have a question, will we see another mobile form of the AIs like the exosuit was? It could be useful to have a drone-like one.

Edited January 6, 2021 by MEDTECH

[Arrow768]

15 minutes ago, MEDTECH said:

But I have a question, will we see another mobile form of the AIs like the exosuit was? It could be useful to have a drone-like one.

It is possible that something like that will be re-added at some point, but it is not planned for the immediate future.

[geeves]

25 minutes ago, MEDTECH said:

But I have a question, will we see another mobile form of the AIs like the exosuit was? It could be useful to have a drone-like one.

speaking as a lone dev and not the team that worked on this whitelisting project, i would be happy to reintroduce it, as it'll be easier to moderate and punish abusers | as always however, we can look at it after the dust settles, since doing a lot of things at once muddles feedback

[Menroka]

The only possible problem I can see are very robust antags, like changelings, because now they are protected under the "Safeguard sapient life on board" law. The AI would have to ensure, that they are captured alive, instead of killed. The AI could make an exception, when a live capture is "infeasible" and would "likely" lead to more dead sapients, however this is ambigous and will lead to problems at some point.

[Skull132]

19 minutes ago, Menroka said:

The only possible problem I can see are very robust antags, like changelings, because now they are protected under the "Safeguard sapient life on board" law. The AI would have to ensure, that they are captured alive, instead of killed.

I don't really see this as an obvious downside. Having antag gamemodes be more complex than "Antag vs station" is a good thing.

[GreenBoi]

1 minute ago, Skull132 said:

I don't really see this as an obvious downside. Having antag gamemodes be more complex than "Antag vs station" is a good thing.

 

22 minutes ago, Menroka said:

The only possible problem I can see are very robust antags, like changelings, because now they are protected under the "Safeguard sapient life on board" law. The AI would have to ensure, that they are captured alive, instead of killed. The AI could make an exception, when a live capture is "infeasible" and would "likely" lead to more dead sapients, however this is ambigous and will lead to problems at some point.

May I suggest borgification, in these trying times?