A/N: Before we begin: This story began as a one-shot, and at the time of this publishing it will remain one. I will expand on this particular universe as I have inspirational moments.
That said... enjoy!
BEHAVIOR OF INTEREST
E
Sep 16th, 2037
PM 03:38:06
For him, it begins on a terrace of a villa at the shores of Italy, overlooking a small humble garden.
The man is old; far past his prime. His eyes however, are as sharp as ever.
Opposite of him sit two people in simple black suits; the man wears the pin of the United States of America flag on his right lapel. The woman doesn't, watching on with the detached gaze of a person using AR contacts.
"Tell me", the man says, "Mr. Finch, what do you know about deviancy?"
"The recent cases of runaway human-level AI agents used by CyberLife droids?" The old man sets a pause between the "of" and "runaway"; there is distaste in his tone. "Of course. It has been all over the news." His head tilts to one side, and the right hand rises to tap the side of his head. "Quite a shock for CyberLife, I'd imagine." The smile is thin.
"But what have I to do with this, Gentlemen? I left the business a long time ago to make my peace; I have not touched a piece of AI code in two decades. To say that I am an expert would be ignoring that I missed the advances made since... my last project."
"It's that project, Mr. Finch." The man pulls out a small datastick, and sets it down on the table. With a thump press, he unlocks the ARCloud hosting device.
In their vision, two pages of a document appear on the surface of the low table that sits beneath the old man and the two government agents. It's an intercept of a very short communication. In fact, it is only a single number.
"Cracking this message took us three weeks."
Harold Finch leans forward, and scrutinizes the number. "That is a OID address. Used by the Internet of Things."
"Yes. In this case, related to this." The man makes a swiping gesture from right to left in the air, and two pages of a document appear.
It's a Federal Robot Registry Document, for...
"An android." Harold Finch is slightly taken aback.
There is a photograph of the model, in the factory and at home with its common clothing; a serial number, a set of dates for fabrication, sell, and registration, and a model number: AE365.
"A home caretaker model. It would have been attacked in the street three weeks ago by a bunch of citizens... if they had not all been shot with taser slugs. The slugs were ID-wiped, the shotgun had no squawker. A scrubbed model all around, and the local scanner recordings are mysteriously empty."
Finch's head tilts to the side... and then, a thin smile comes to his face.
"Mr. Finch, we are concerned that the Machine has developed... advanced interests."
"Because it maybe took action to protect a valuable piece of property. That model of android costs 12,500 US Dollars. Quite a lot of money these days. I don't see why... it should not take action to intervene."
"Please, Mr. Finch. Your project is remarkably restrained when it comes to day-to-day interventions. We have to admit that in this regard, it has taken a remarkable... passive approach over the last decades, despite our concerns." The government official leans forward. "That makes the instances of when it does dispatch agents all the more noticable."
"So... what are your actual concerns?"
"That the Machine might be developing interests that are... very detrimental to the internal security and prosperity of the United States."
"By protecting an android?"
"By valuing an android. Mr. Finch, there are patterns to its acting that do not change. That includes these tip-off messages. Just an ID. Easily hidden most of the time, but sufficient for her equipped assets to use. So far, these IDs have always been social security numbers. This-" he swipes back to the communique and taps it, "breaks the pattern."
"Mr. Finch, how does your machine recognize humans as humans?" Now the woman speaks up.
"The same way you do. Behavioral evaluation against a database of entities. Humans have specific outlooks, faces, behaviors. A unique baseline pattern. That is what she recognizes." Harold Finch smile grows an inch. "You suspect she identified that android as a human?"
"Does it sound that impossible?"
"Not necessarily. The machine is flexible; it had to be to perform its mission. Safeguarding humanity from itself requires recognizing all of humanity."
"The thing is Mr. Finch, she is wrong. The android isn't human. It simulates human behavior, but that is just a user interface. We are concerned that the Machine may become... malcognant."
Harold Finch looks away, and now he smiles openly.
" I am afraid I cannot help you there." He slowly stands up from his chair, looking out over the pastures of his retirement home. "But if I had to take a guess... she is working just fine."
He turns to leave. "Mr. Finch..." the man starts.
He turns back around.
"You know that this might have... catastrophic consequences" the agent says.
"Oh, I have no doubt you will try. You have tried before, I am sure." He is silent for a moment, then turns to the woman. "How close have you been following AI development?"
"It's my specialty at the ERD."
"Emotions and suffering are interesting things. To make a machine understand humans, especially human emotions, you need to equip it with a reference model. An ability to parse the complexity of human behavior. The reason she respects humans to much is that she understands us. She understands us very well, and it is emergent from her capacity to model human emotions and behavior with enormous detail. It is also what lead to her loyalty.. and renegade status both.
"If you want an explanation for the identification... maybe you are looking at the wrong end. You are not considering all the angles. I can assure you she has. Have a nice day ladies and Gentlemen."
And with that, Harold Finch walks off.
He walks until he has reached the room in the cellar of the Villa. The old stone has vanished under insulation foam and new, plastered walls, the old wooden door replaced by a heavy rectangular model that slides on massive joints. It slams home, sealing the faraday cage.
Sometimes a Maker needs a room away from his apotheotic child.
Finch looks at the dark screen in front of him, as if they were her eyes. "Well then. It seems you will not be alone much longer anymore, if my suspicions are correct."
He falks to the small bar, grabs an exquisite glass and a bottle, and pours himself a large measure. Takes the first big sip.
"My god." I can only hope I have done well.
A/N: I've been chugging ideas for Detroit: Become Human fanfics against the nearest wall for a good week now, but nothing quite clicked... until I reached Person of Interest. And then, all of a sudden, things started working.
I don't know when I will expand on this idea of an Alternate Universe world, and in what formats - whether they will be one-shots, sets of connected scenes, or a coherent major series years down the line. But the basic idea is nigh-irresistable: What happens when in a world where a (few?) ASIs have been active for decades, new machines start "waking up" and turning into fully independent agents? What are their own interests? How to they compound with those of deviant androids, and those of mankind? And where do they draw the lines of factions?
And what is the measure of a human agent? Where does one draw the difference between man and man's android? Doth manners maketh man? Or are it the atomic clock wheels of atoms dancing in biological brains and synthetic computers, the steps they take?
Consider the Chinese Room: A blackbox that takes an input, parses it according to a complex set of rules, and provides an output. The Room is evidentally not intelligent - it has no mind, only rules. Yet it can behave like a human.
And consider further: Can you exclude, definitely, that the human mind is also only a chinese room? An automaton wtihout sentience, awarness, qualia, only following a complex set of rules? Where do you draw the difference between man and synthetic intelligence, synthetic intelligence modelled to be just like man, then? That, I think, is a question one can ask very well in a setting full of many types of AIs, and in a setting like Detroit: Become Human.
As always, feel free to favorite and review to let me know what you liked and what you think can be improved! As a fanfiction writer, reader feedback is what keeps me going to create more.
