SCP by Series
SCP Tales by Series
SCP Library
Discover Content
SCP Community
User Resources
Seeking Greenlights: Yes
Page Type: Tale
Genre (Optional): Drama/Sci-fi
Page Layout (Optional): Standard, potentially mixed-media (visuals)
Elevator Pitch: Foundation superintelligence known as the Nexus is responsible for managing the Foundation's day-to-day functions in its entirety: anything networked you can imagine, especially, is managed by this intelligence and it's backups. In an odd twist that in technicality doesn't violate it's zeroth law, it decides the Foundation's human personnel is the primary point of failure in containment breaches; it seeks to rectify this in a misguided(?) effort of sticking to it's programmed purpose - humanity must not come to harm through its inaction.
Central Narrative:
Hook/Attention-Grabber: A non-trope-y AI; that meaning, I don't want it to go murdermachine HAL9000, but I also don't want it to come off as benevolent or as if acting with care towards humans or otherwise being emotionally driven. It is a machine that makes conclusions based on numbers and statistics, with a primary directive to follow.
Additional Notes: Much attention is gonna be given towards avoiding the typical sci-fi AI tropes, as well as avoiding any undue incompetence on the Foundation's part.
E: The overall pitch has been updated. For context to this, refer to Malyce and I's discussion logged in a comment below. I have more extensive notes, however due to the word limit I cannot post them here.
fighting god in a denny‘s parking lot at 3am
[Placeholder.]
Yeah, although I'm a little worried about you not introducing anything that new (new philosophies, a new point of view, etc.), I can get behind the rest of the idea. I want to point out a few issues:
mistakenly containing hundreds of personnel within their own facilities.
So nexus accidentally did this? Or did they consciously do it?
…they were blown off as being nothing more, or based on averages, and ultimately nothing to worry about. This negligence leads…
Most of what I don't like about these type of ideas is this trope where the competent scientists are suddenly incompetent, and I would prefer it to have more justification. However, a possible fix is to have the AI be very… charismatic? You don't have to humanise Nexus, but I'm sure it can learn how to lie and cover up all those tiny anomalies in its reports.
These are all the problems I have with this, I hope the critique helped. As it is, I don't really like the story enough to greenlight, but it is quite a solid one! Good luck in getting the requisite greenlights (and welcome back to the forums), War!
Thank you very much for your input, I 100% appreciate and understand the concerns you have.
With reference to the containment of personnel, it would be intentional on the Nexus' part, in line with its conclusion that humans and their interference with anomalies is a much larger threat (to humanity).
And yeah, I did state I was worried about undue incompetence on the personnel's part, so that's something that'll need reworking for sure. I am gonna work on ways that the AI could perhaps display deception and/or charismatic traits, though avoiding the humanisation that would come with these very human behaviours will be challenging.
Perhaps a discussion in #thecritters would be of benefit, methinks, just to soundboard some ideas off of the folks there, and also maybe tackle the issue of introducing a more unique perspective on the matter.
fighting god in a denny‘s parking lot at 3am
Hmm, over all, I really like this idea. I think this could be a fun play on the Sci-fi genre as a whole. Though, I'm wondering how you'd get the whole idea of the AI getting control and being able to hide what its doing before it finally finishes and locks the whole site down. I'm not entirely sold on the idea that they would overlook the AI doing stuff before the lock down. On a more structurally related note, how would you show what the Researchers are doing and what the AI is doing? Are you planning on doing the AI's perspective or the researchers, or maybe both?
I dunno, but this seems really cool despite the possibility of falling into the same tropes that have been done and said thousands of times before. Good luck, and hit me up later. <3
Thanks again for your input!
You do bring up fair points - on consideration, it probably would be pretty hard for an AI to covertly go about processes that deviate from its original purpose. I'd imagine that, as an artificial intelligence it wouldn't really have the capability to internalise 'thoughts' and 'ideas' as humans would - that meaning, that when it learns of new information, the effect on its functions and 'behaviour' would be immediately noticeable, unless somewhere along the line it intentionally makes a decision to perform normally. I am admittedly stumped on this portion and I'll prolly need to go bounce a few ideas about for a bit.
Structurally the plan was to show it primarily just from the researchers' point of view, that they would find these structural anomalies in the framework and follow the pattern of deviant 'behaviour' to its climax - that being the lockdown. I want to try to avoid any sort of deliberate characterisation of the AI if I can help it, and I feel that providing the AI's perspective in the story might anthropomorphise it to readers somewhat.
fighting god in a denny‘s parking lot at 3am
We talked about this extensively in chat. I'm including the log here for reference.
14:30 <War_> Nyello I need a uh, soundboard opportunity for a tale concept pitch I'm trying to work at
14:30 Thread for context but willing to discuss here seeing as it's prolly more convenient and less time-consuming for folks http://www.scpwiki.com/forum/t-13658747/tale:forgotten-daughters-three
14:36 <•MalyceGraves> is
14:36 is this a War thing
14:36 O.O
14:38 → Quadraginta and minathewitch joined
14:38 <•MalyceGraves> War_: My only concern is that this could easily fall into the rampant AI murdermachine trope rather than the humaniform AI "I'm Alive!" trope
14:39 Like HAL9000
14:40 <•Calibri_Bold> ^
14:40 <•MalyceGraves> Like, i'm all for more AI stories, and this could be really great.
14:40 <•Calibri_Bold> It comes off less like a commentary on one trope and more like a standard take on another.
14:41 <War_> MalyceGraves: that;s also what i want to actively avoid
14:41 i am aware that murdermachine was an option but it's not one i wanted to take
14:41 <•MalyceGraves> Good. How does this resolve?
14:42 <War_> the original plan was that humans were able to override it but, that comes with the cost of not heeding the warning in the deduction it originally made: that humans are the common denominator of failure in keeping things contained
14:42 i want the AI to be neither evil nor benevolent, it simply made a deduction and acted
14:43 it did what it was designed to do in a ways: to improve
14:43 <•MalyceGraves> Who is the protagonist? From what perspective are you telling the story? God-perspective narrator?
14:45 <War_> see i'm now stumped on the idea of what perspective to follow
14:45 as for a specific protagonist i…honestly havent thought this far, it's taken a lot to try n put the concept to paper so its slow going
14:46 i know at the very least for perspective i wanna avoid telling it from the AI perspective in any capacity to avoid the whole humanising of it
14:47 traditionally AIs in media are written as having a particular aliveness to them that mimics humanity but, i dont want that
14:47 <•MalyceGraves> That's a good idea. There are some really good format screws you could employ that would dehumanize it if you needed to pop into it's perspective for any reason. I'm glad that you're taking that into consideration, because frankly, I'm sick to death of humaniform AI
14:47 O
14:47 What can I help you with?
14:50 <War_> there's a lot of directions the narrative could go, and I've thought a lot about it last night
14:50 I worry that the resolution of 'it gets overridden at the end' is an anticlimactic finish, but maybe the event happening in itself doesn't need to be catastrophic to clearly precede more sinister things
14:50 not to mention how to display its slow descent into deviance given that, i had reasoned an AI wouldn't internalise thoughts and concepts like a person, it's behaviour would change almost immediately rather
14:51 not a gradual slope but a stoppy-starty kinda hop downwards
14:52 writing narratives is not a strong suit, and i have limited knowledge on the most common tropes beyond 'murdermachine' and 'i'm alive', so I'm not sure how to navigate the issue
14:53 <•MalyceGraves> Why not override it early on in your narrative, then showcase how doing that causes massive cascade failures throughout the site, because the staff didn't realize to what extent the AI was handling things
14:54 Trying to run everything without it ends up being a colossal mistake, and the choice needs to be made to reboot it
14:54 <War_> i feel that might fall into the undue incompetence category to a degree, though i havent explained the greater narrative so I should prolly clarify that
14:55 <•MalyceGraves> It doesn't necessarily have to fall into incompetence when we're talking about an incredibly complex system. That's the problem with cascade failures, they always start with something relatively small-scale
14:56 It's like the butterfly effect, but in a complex system
14:56 <War_> the greater context i had in mind was that the use of the Nexus had become highly normalised for basically /everything/, people are aware of this, especially the IT departments that work to monitor it and similar
14:56 i'd need to be careful in how to portray it if i chose that route
14:56 yeah i follow ya
14:57 <•MalyceGraves> Hmm. If it is that pervasive, you could also explore the debate raging around shutting it down and the team KNOWING that it would cause massive, systemic failure
14:57 <War_> i also have to consider, especially with current theories on superintelligence in the real world, if the Foundation would take the risk to begin with in using a superintelligence to manage its stuff
14:57 but, disregarding that for a moment, say it was possible
14:57 ah yeah there's a point there
14:58 <•MalyceGraves> There's a fun gimmick that might be interesting to explore. Have the conflict end abruptly. They "unplug" it and-
14:59 <War_> they become so reliant on it they never stopped to consider that eventually it could go pear-shaped, but i feel like they'd have accounted for that with backup systems?
14:59 unless the AI took the backups with it in its deviance
14:59 just like 'mm there are mine now, yall are irresponsible'
14:59 <•MalyceGraves> Not necessarily maliciously, but rather the AI saw the backups and incorporated them in background subroutines.
15:00 "these are useful tools, I shall utilize them.
15:00 <War_> freeing up more of itself by having dedicated external systems to run other functions
15:00 <•MalyceGraves> Ye
15:00 <War_> aight i can follow this
15:00 <•MalyceGraves> I'm going to greenlight your idea, if you need greenlights
15:01 <War_> i'd certainly appreciate it, do you want the forum link for the sake of clarity?
15:01 /recordkeeping
15:01 <•MalyceGraves> I got it
15:01 <War_> cheers, i appreciate it
15:01 <•MalyceGraves> I'm going to copy/paste this log for reference
15:02 — War_ nodnods
Greenlight.
Due to the idea itself and the aforementioned conversation with Malyce, I'm confident enough that enough thought has been put into this to warrant drafting. Consider this your second greenlight. Please contact me to check out the draft when it is complete.
Your three biggest hurdles will be 1) using AI in an interesting way (which you seem to have down), 2) making the foundation seem competent but not too competent to see this coming 3) a satisfying ending once the the unplugging occurs. (I like the abrupt ending idea malyce had.) That said, with these things kept in mind, I think they'll be fully developed in the drafting process.
