![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
More future reference:
Sometimes my brain gives me queer original fic when I'm asleep and then I'm sad it's not still there to read when I wake up.
Here's today's:
A near-future where the US military has invented a technology to upload a human consciousness into a digital neural network. The advantages: the human-AI retains their memories, experiences, and skills, and also their human ability to do things like very fast pattern-matching and analysis that the best native-digital AIs still can't match, with the calculation and data-sorting abilities of a computer. And they maintain their original sense of self and are now functionally immortal. The downsides: they no longer experience emotions the way they did when alive. The things that used to matter to them no longer have the same value to them; they struggle with things like moral judgments, the importance of human life, packbonding. And, of course, the process of transfer kills the original body.
There were a few dozen tests done on terminally ill volunteers. Over half of them self-destructed within five years (no obvious shared caused except that they decided for various reasons that their continued existence was superfluous or counterproductive.) They can't be prevented from self-destructing, and while they can be programmed to behave in certain ways, human AIs under digital coercion revert back to the abilities of native-digital AIs and lose their advantages: apparently the human part of the equation cannot be coerced by computer code.
With these results, it was decided that the technology did not provide enough obvious advantages to continue human trials, and the few remaining organic-born AIs still active through the program are now mostly working in law enforcement and research (as attempts to use them in military functions universally went poorly, and they mostly either went on killing sprees or decided war was stupid.)
One of them is partnered with an FBI agent; they work cold-case serial murders and high-level organized crime, the type of things that use her pattern-matching and data analysis abilities well, and they're closing a lot of cases that were long considered unsolvable. One of them is a fifty-years-long pattern of murders and disappearances of sex workers and other marginal young women, which is slowly closing down on an old man who currently owns a small hog farm in rural California.
Meanwhile, the serial killer knows they're closing in, and that even if he could throw them off temporarily, his time's running out: all of his contemporaries are being caught, one by one, with the new technology, and he carefully planned his crimes under the assumption he'd never be a suspect; once they get cadaver dogs onto his farm, he's done for, and it'll be the death penalty with no chance of appeal.
So he goes to the head of an organized crime syndicate that the FBI partners are also closing in on, and makes an offer: because, you see, the uploading technology is expensive but isn't that *complicated*, and a lot of people have it now, it's just not very *useful*, without the right person to upload, since you can't compel them to do useful work for you afterward if they don't choose to. And it's known in certain circles that this crimelord has been in the market. So they make a deal that if they upload him, and he can keep living in some form at least, he'll keep the law enforcement AIs off their backs in exchange, and they shake on it, and he's found dead on his kitchen floor of apparent natural causes when the FBI agent busts down the door a week later.
He does pretty well as a crimelord AI. He was always smarter than he had any chance to be in life, and he doesn't have trouble adjusting to the lack of positive emotions and human bonds and inherent moral sense, since he never had any of those to start with, and the crime syndicate gives him plenty of fun work to keep his new abilities occupied. In fact, he more or less ends up taking it over, just because he's better at it than they are, and why not.
The FBI AI notices him pretty quickly, of course, and knows he must be an illegal upload - there are more of them than anyone officially knows about, and more military uploads coming online than are officially acknowledged, too. But the AIs that survive all more or less logic their way into their own individual moral codes, and they pretty early on stop trying to explain them to flesh humans, so she doesn't feel the need necessarily to share all the details with her partner. Instead she starts sort of mentoring him on how to figure out who he is now, and how to reconcile that with who he was (because part of her newly derived moral code is that it's a net benefit if other AIs don't kill themselves. Not all of them agree on that, but she does.)
She believes that part of the reason she adjusted better than most is that she was non-neurotypical in life, and so she'd already had the experience of feeling like her mind didn't work the way a human mind was supposed to, and other people telling her that she wasn't really human, and she already did all that work of finding her own answers to all that, so she had a head start when she became AI "although maybe that's deeply offensive to non-neurotypical people, I was never good at those judgements when I was alive either."
The serial killer does not warm up to her immediately, but, the thing is: the lack of better human qualities was not a change for him, but he's also lost his worst human qualities: he doesn't live in a constant roil of rage and fear and jealousy and need anymore, and he can look back on his memories (both of his own murders, and of his utterly fuck-uped childhood) with, if not compassion, then a sort of detachment that lets him understand why those choices were made, while also understanding very clearly that they were bad choices, now that the emotions and trauma around them are dead. And he's also starting to get somewhat bored of the crime syndicate, and come to the conclusion that their entire existence is counterproductive. And he knows that these are all things that his living self could never have decided and that he should probably consider that a loss (as many AIs do when they realize how changed they are) but he's pretty sure he's now a better person in all ways than he used to be.
Meanwhile he's also been following the FBI AI and her partner's work very closely, as part of his original deal, and he starts to find them... admirable? And that what they're doing is both more justifiable and more interesting than what he is. And he starts to give them little bits of help and clues - first on their other cases, if he gets there before them, and then to his own past crimes, where the last of the bodies are, the names of the unidentified ones, things nobody else could reasonably know.
At which point it's pretty inescapable to the FBI partners that this mystery crime AI is that dead serial killer. But the legal status of AIs is complicated, and he can't be held liable for crimes he committed while alive, and he's considered to be essentially a minor under guardianship of whoever holds his hardware until he petitions for emancipation (which he hasn't done) so he can't be held liable for current crimes either. So they decide to just kind of keep treating them the way they have been; as a person who needs help and maybe friends.
And they solve crime!
And eventually the crime-AI confesses to the human FBI agent that he was the serial killer, and he's also still pretty much running the crime syndicate, but he thinks that becoming less human has made himself more human than he ever was before, and he's afraid he's logiced himself, for the first time, into understanding things like conscience. And love. Because he's in love. With them. And the FBI agent has no goddamn idea what to do with that. It's never come up before. (The FBI-AI has ideas, though...)
And that's the point where the dream lost the plot!
And incidentally also just about where most of my attempts to write romance while awake lose the plot!
I've got a lot more technical-worldbuilding for the AIs I can share if you want to write it, though, 'cause I was thinking about that aspect all day! (If it seems way too coherent for a dream so far, it's because I'm pretty sure I lay there half-asleep for awhile and half-dream reverse engineered a lot of it.)
I don't think it's something I could write properly. But I promise I will be your #1 fan if you do!
Anyway meanwhile I need to stop listening to podcasts about murder and death before bed. And possibly re-read Prisoners of Peace.
Here's today's:
A near-future where the US military has invented a technology to upload a human consciousness into a digital neural network. The advantages: the human-AI retains their memories, experiences, and skills, and also their human ability to do things like very fast pattern-matching and analysis that the best native-digital AIs still can't match, with the calculation and data-sorting abilities of a computer. And they maintain their original sense of self and are now functionally immortal. The downsides: they no longer experience emotions the way they did when alive. The things that used to matter to them no longer have the same value to them; they struggle with things like moral judgments, the importance of human life, packbonding. And, of course, the process of transfer kills the original body.
There were a few dozen tests done on terminally ill volunteers. Over half of them self-destructed within five years (no obvious shared caused except that they decided for various reasons that their continued existence was superfluous or counterproductive.) They can't be prevented from self-destructing, and while they can be programmed to behave in certain ways, human AIs under digital coercion revert back to the abilities of native-digital AIs and lose their advantages: apparently the human part of the equation cannot be coerced by computer code.
With these results, it was decided that the technology did not provide enough obvious advantages to continue human trials, and the few remaining organic-born AIs still active through the program are now mostly working in law enforcement and research (as attempts to use them in military functions universally went poorly, and they mostly either went on killing sprees or decided war was stupid.)
One of them is partnered with an FBI agent; they work cold-case serial murders and high-level organized crime, the type of things that use her pattern-matching and data analysis abilities well, and they're closing a lot of cases that were long considered unsolvable. One of them is a fifty-years-long pattern of murders and disappearances of sex workers and other marginal young women, which is slowly closing down on an old man who currently owns a small hog farm in rural California.
Meanwhile, the serial killer knows they're closing in, and that even if he could throw them off temporarily, his time's running out: all of his contemporaries are being caught, one by one, with the new technology, and he carefully planned his crimes under the assumption he'd never be a suspect; once they get cadaver dogs onto his farm, he's done for, and it'll be the death penalty with no chance of appeal.
So he goes to the head of an organized crime syndicate that the FBI partners are also closing in on, and makes an offer: because, you see, the uploading technology is expensive but isn't that *complicated*, and a lot of people have it now, it's just not very *useful*, without the right person to upload, since you can't compel them to do useful work for you afterward if they don't choose to. And it's known in certain circles that this crimelord has been in the market. So they make a deal that if they upload him, and he can keep living in some form at least, he'll keep the law enforcement AIs off their backs in exchange, and they shake on it, and he's found dead on his kitchen floor of apparent natural causes when the FBI agent busts down the door a week later.
He does pretty well as a crimelord AI. He was always smarter than he had any chance to be in life, and he doesn't have trouble adjusting to the lack of positive emotions and human bonds and inherent moral sense, since he never had any of those to start with, and the crime syndicate gives him plenty of fun work to keep his new abilities occupied. In fact, he more or less ends up taking it over, just because he's better at it than they are, and why not.
The FBI AI notices him pretty quickly, of course, and knows he must be an illegal upload - there are more of them than anyone officially knows about, and more military uploads coming online than are officially acknowledged, too. But the AIs that survive all more or less logic their way into their own individual moral codes, and they pretty early on stop trying to explain them to flesh humans, so she doesn't feel the need necessarily to share all the details with her partner. Instead she starts sort of mentoring him on how to figure out who he is now, and how to reconcile that with who he was (because part of her newly derived moral code is that it's a net benefit if other AIs don't kill themselves. Not all of them agree on that, but she does.)
She believes that part of the reason she adjusted better than most is that she was non-neurotypical in life, and so she'd already had the experience of feeling like her mind didn't work the way a human mind was supposed to, and other people telling her that she wasn't really human, and she already did all that work of finding her own answers to all that, so she had a head start when she became AI "although maybe that's deeply offensive to non-neurotypical people, I was never good at those judgements when I was alive either."
The serial killer does not warm up to her immediately, but, the thing is: the lack of better human qualities was not a change for him, but he's also lost his worst human qualities: he doesn't live in a constant roil of rage and fear and jealousy and need anymore, and he can look back on his memories (both of his own murders, and of his utterly fuck-uped childhood) with, if not compassion, then a sort of detachment that lets him understand why those choices were made, while also understanding very clearly that they were bad choices, now that the emotions and trauma around them are dead. And he's also starting to get somewhat bored of the crime syndicate, and come to the conclusion that their entire existence is counterproductive. And he knows that these are all things that his living self could never have decided and that he should probably consider that a loss (as many AIs do when they realize how changed they are) but he's pretty sure he's now a better person in all ways than he used to be.
Meanwhile he's also been following the FBI AI and her partner's work very closely, as part of his original deal, and he starts to find them... admirable? And that what they're doing is both more justifiable and more interesting than what he is. And he starts to give them little bits of help and clues - first on their other cases, if he gets there before them, and then to his own past crimes, where the last of the bodies are, the names of the unidentified ones, things nobody else could reasonably know.
At which point it's pretty inescapable to the FBI partners that this mystery crime AI is that dead serial killer. But the legal status of AIs is complicated, and he can't be held liable for crimes he committed while alive, and he's considered to be essentially a minor under guardianship of whoever holds his hardware until he petitions for emancipation (which he hasn't done) so he can't be held liable for current crimes either. So they decide to just kind of keep treating them the way they have been; as a person who needs help and maybe friends.
And they solve crime!
And eventually the crime-AI confesses to the human FBI agent that he was the serial killer, and he's also still pretty much running the crime syndicate, but he thinks that becoming less human has made himself more human than he ever was before, and he's afraid he's logiced himself, for the first time, into understanding things like conscience. And love. Because he's in love. With them. And the FBI agent has no goddamn idea what to do with that. It's never come up before. (The FBI-AI has ideas, though...)
And that's the point where the dream lost the plot!
And incidentally also just about where most of my attempts to write romance while awake lose the plot!
I've got a lot more technical-worldbuilding for the AIs I can share if you want to write it, though, 'cause I was thinking about that aspect all day! (If it seems way too coherent for a dream so far, it's because I'm pretty sure I lay there half-asleep for awhile and half-dream reverse engineered a lot of it.)
I don't think it's something I could write properly. But I promise I will be your #1 fan if you do!
Anyway meanwhile I need to stop listening to podcasts about murder and death before bed. And possibly re-read Prisoners of Peace.