An examination of A.I. and E.I. using Ex Machina and other movies as a reference.

OldHat63 · Sep 19, 2018

First off, let me start by saying there's already another thread here concerning writing a more believable A.I.

It can be found here: https://absolutewrite.com/forums/showthread.php?338260-Writing-Believable-Artificial-Intelligence

This thread is not intended to be that, but more of an examination of what has been presented so far in science fiction, and what I personally belive the writers have gotten correct, what they've gotten wrong, as well as what they've either glossed over or simply ignored.

I hope these things will at least be of interest to anyone writing on the subject, whether it's for fictional purposes, or simply covering some real-life event or technical article. And yes, it's probably a bit arrogant of me to think I have any business commenting on such, since I have no degree in robotics, programming, neurology, or other associated field.
( I do have a technical degree that involves mechanical design, and have spent many years working on and with computers. I've even built a couple of the infernal things. Also I've spent most of my life studying the human body and mind. That's mostly having an interest in drawing and martial arts, as well as living with crazy people. )

Now that's out of the way, I'll get to the subject at hand.

One, I have no small amount of problem with the term "Artificial Intelligence". Mostly because it's not really what people mean when they use it. They are actually thinking of a real, couscous, thinking mind, not something that just appears to be one, at first glance. And no, I don't think this is just a matter of semantics, or a misunderstanding. I'm fairly certain that most people simply believe that "real" intelligence is and has to be, by definition, organic in nature.

Why do I say that? Because the very word "artificial" means something that appears to be something else, only bears a superficial resemblance to that thing, but contains none of it's actual properties. It's a fake. An imposter that cannot and does not function in the same way that what it's mimicking does. A plastic or silk flower may look like the real thing. It can even be manufactured to feel or smell like the real thing. But it won't have pollen, won't have nectar, and won't grow or wilt. It serves only as a decoration.

The same is true of an artificial intelligence - which I believe we already have plenty of. It/they can give responses to questions, but only so long as those responses fall into a category that it is already programmed with, or has been allowed and directed to self-program with. It won't truly understand either the question, or the answer though. And the answer won't be one of it's choosing, so much as it will be the closest match for a pre-determined set of qualifications. A if/if not list, as it were.

A human mind doesn't work that way. If a person is asked a question about something they have no experience with, or know very little about, they may say as much, or they may guess, based on something else they do know. A computer program can't do that. At least not yet. And it can't just make something up and bluff it's way through either. People can and do that very thing.

In Ex Machina, you can actually see the machine learning, through the various itinerations of android that Nathan has built before Eve, in the flashback videos. One is even aware that it's a prisoner and wants out. And later on, Eve and Kyoko plot and conspire to escape.

Sorry, folks, but that's not an "artificial" anything; That's a working, thinking mind, even if it's a constructed or engineered one, housed in a synthetic gel.

And speaking of that, here's something the movie touched on, and I think got very right, but didn't really elaborate on too much: the mechanics of an artificial brain, and why I doubt any computer like we have today will ever achieve any sort of real intelligence, no matter how complex the program is. Here is a section taken straight from the movie script:

( Nathan : )

"Here we have her mind. Structured gel."

The axon-like tendrils glitter and flicker with tiny pulses of light.

( Nathan : )

"Had to get away from circuitry. Needed something that could arrange and rearrange on a molecular level, but keep it's form where required. Holding for memories, shifting for thoughts."

So, the writers recognized that "hardware" wouldn't cut it, so they used "wetware" instead. Sort'a like we humans do, no?

Where they completely dropped the ball is in not telling us how Nathan programmed this thing, only stating that the "software" was his search engine... his version of Google. No mention at all of the fact the usual ones and zeros wouldn't cut it, and that if he wanted to use any normal programming language, he'd be need ing some other program to work as a translator, of sorts, to move that gel around into the appropriate configurations.

...or that the very same technique would likely work on us humans with not much modification.

A movie a year earlier, "The Machine", did pretty much the same thing, though like "Terminator" before it, really didn't touch on the mechanics of the mind, sticking strictly to the whole "We figured out to record and download a whole human personality while we were working on this sucker". Like it really was just a matter of the right program code, even though their machine looked VERY MUCH like a human with a fiber-optic nervous system when she got going... or maybe a bio-luminescent one?

Hard to tell, and the writers didn't say, for obvious reasons.

And then they went right back to the same old, "it's just a program" crap, at the end, with the "child" that was basically stuck in a tablet as the human "dad" and android "mom, looked out over the ocean. Leaving the audience without any real explanation of how, or if it was really anything more than just a copy of a dead kid's engrams, used as a template for a program in a remote computer installation somewhere to "animate".

And Just as Ex Machina ended with Eve standing at the crosswalk, watching people, that was the end of the movie, "The Machine"... and should probably be the end of this post as well, for now.

As I said, these are just a few of my observations and thoughts on the subject. If you have any of your own, from either these movies, or some other source(s), pull up a chair and tell us about 'em.

O.H.

onesecondglance · Sep 19, 2018

I'm not sure there's any usefulness in making a semantic distinction between "artificial" and "constructed" or "engineered", but taking you at your word that the filmmakers got these things wrong, how might you improve these films?

lizmonster · Sep 19, 2018

onesecondglance said:
I'm not sure there's any usefulness in making a semantic distinction between "artificial" and "constructed" or "engineered", but taking you at your word that the filmmakers got these things wrong, how might you improve these films?

The main thing that would've improved Ex Machina for me was if I hadn't bothered to sit through it.

I found that film a pretty tired retread of a lot of old ideas and unpleasant stereotypes.

I agree, though, that how the intelligence is created isn't really relevant. I'd also argue that whether or not it's genuinely independent or just following its programming is a moot point. (We don't even know that about the human brain; there's a good argument to be made that there's no such thing as free will at all.)

There's a much-lauded episode of ST:TNG: "The Measure of a Man." In it, everyone's favorite yellow-eyed android is subject to a trial (of sorts) to prove his sentience. What's always bugged me about the episode is that there's no real conclusion here: Picard makes a pretty speech (as he so often does), and the judge is persuaded. There's no scientific argument presented at all.

But we the audience are pleased by the conclusion, because we like Data. Our perception of his behavior makes him indistinguishable from any biological sentient creature. And that's exactly why the judge's conclusion is correct (whether or not the script makes a convincing case, and I really think it doesn't): if we can't tell the difference, we must assume sentience. To behave otherwise would be morally abhorrent.

So...I think whether or not an AI (or EI or whatever term you choose to define) is a genuinely sentient mind or not is the wrong question. The question is, if we become capable of producing something that is in every way indistinguishable from a human mind, what are our moral and ethical obligations to this creation?

To me, it's less about how we define the created intelligence (and in SFF, we can define that however we want), and more about how we define humanity.

bwebs · Sep 19, 2018

lizmonster said:
The main thing that would've improved Ex Machina for me was if I hadn't bothered to sit through it. I found that film a pretty tired retread of a lot of old ideas and unpleasant stereotypes.

Awww...I really liked Ex Machina

I have no idea what was intended, *spoilers* but for me I really liked the billionaire who has everything, confronted by meaninglessness, attempting to construct his own great pyramid (i.e. Ozymandias) only to find more meaninglessness inside--and what that would do to a person, especially isolated, especially contrasted with the naive worker confronting the human bias toward anthropomorphism and then when he cuts open his own arm seeing how anthropomorphism and meaninglessness are two sides of the same coin. I even liked the ending thinking that even though the AI is loose now in the world it still hasn't escaped the exact same dilemma. But maybe that's just me

I also like the film making.

In a similar vein it makes me think about the AI in Ian Banks Culture novels. The ship minds, exist in most ways above humans on the social hierarchy, are much superior etc, but if I remember right are modeled after human brains. The AI's themselves have tried to build "pure" AI's not modeled on biological life, but when they do, those pure AI's always "sublime", which is sort of like move on to an afterlife/oblivion--no one really knows. This vexes the human modeled AI's to no end, the entire "culture" in this universe having a sort of jealous fixation on the sublimed, but don't want to do it themselves. Sort of like having an ascetic monk attending your party in the Hamptons and be unimpressed.

Harlequin · Sep 19, 2018

"Intelligence" is a synonym for "humanness", no more and no less. This is why (gasp of surprise) we are the most intelligent animals on the planet; the criteria is defined on human traits and abilities, with ourselves as the measuring stick.

When we talk about creating artificial intelligence, we are really talking about creating artificial humans; so in that sense, I think it is perfectly appropriate.

lizmonster · Sep 19, 2018

bwebs said:
Awww...I really liked Ex Machina

Oh, I'm aware I'm an outlier on that one.

There are three other films on my list of Movies That Actually Enraged Me, and two of them won Best Picture Oscars. Stuff that pisses me off doesn't have much to do with craft.

bwebs said:
In a similar vein it makes me think about the AI in Ian Banks Culture novels. The ship minds, exist in most ways above humans on the social hierarchy, are much superior etc, but if I remember right are modeled after human brains. The AI's themselves have tried to build "pure" AI's not modeled on biological life, but when they do, those pure AI's always "sublime", which is sort of like move on to an afterlife/oblivion--no one really knows. This vexes the human modeled AI's to no end, the entire "culture" in this universe having a sort of jealous fixation on the sublimed, but don't want to do it themselves. Sort of like having an ascetic monk attending your party in the Hamptons and be unimpressed.

I haven't read Banks, but Spouse has read most of his work and we've spoken about it. I quite like the idea of non-human AIs - designed by other AIs, or by aliens - as a story point. It's always fascinating to explore the idea of what "intelligence" really means outside of human context.

Of course, since we can't really get out of human context, even our wildest imaginations are limited.

bwebs · Sep 19, 2018

Lizmonster--as far as non-human context, or even just different human context goes, I really like Banks. Re: stuff that pisses me off randomly, I can't tell you how often I vindictively start rooting for protagonists everyone else seems to love to get run over by a bus

Harlequin--not arguing per say, but my understanding of general intelligence is that it is a matter of pattern recognition regardless of context, which I always thought of as more of a factor of pure processor speed and little to do with programming (humanness). I'm going to think about that a bit though!

Harlequin · Sep 19, 2018

Sure, and it's human centric as a definition.

Intelligence is multifaceted, and relative to a given sample population. But conceptually it is grounded in, and determined by, human perception & human norms.

It has to be, because no other species has proposed any other model, so only our perspective informs the subject. And we cannot think or perceive very far outside our own parameters.

OldHat63 · Sep 19, 2018

onesecondglance said:
I'm not sure there's any usefulness in making a semantic distinction between "artificial" and "constructed" or "engineered", but taking you at your word that the filmmakers got these things wrong, how might you improve these films?

Well, you could start by making humans smarter and less egotistical and narcissistic.

I mean, from what I've seen, if we're the best the universe can come up with, not only is it a terrible waste of space, it's pretty boring and unimaginative as well. :Shrug:

Past that, I dunno... I tend to like the more technical stuff, even if it's only hypothetical, or just a logic-based guess. But I suppose that stuff doesn't really fit in with "entertainment" for most people.

The philosophical end of it... the "okay, we can, but should we?" also has an appeal. But probably falls into the aforementioned category.

...So I don't know. I did like the movie. I just wanted... more.

O.H.

MaeZe · Sep 19, 2018

My somewhat of an educated opinion:

You might as well go with any fiction that suits the story. Until we understand the nature of our own consciousness, it's doubtful one could recreate it artificially.

Faux AI is essentially lots of memory (even learning new information), a large active memory (is it still called RAM?) and very complex algorithms.

How do I know because obviously I'm not tech savvy? Because I know a lot about how the brain and consciousness work. Consciousness is only one thing going on in our brains that collectively ends up being us. We've learned a lot about this from studying brain injuries.

Take typing, most of us are pretty fast at that. Do you ever catch yourself typing a different word than is in your conscious thought? And I don't mean spellcheck, I mean catch yourself in the act. You think 'your' and your fingers have already typed 'you' and moved on to the next word. Just now as I typed the above, I consciously thought 'to' but my fingers typed 'the'. How often have your typing fingers gotten ahead of your thoughts?

That's one of the simplest examples. Obviously we think nothing of the fact the brain is controlling all sorts of body functions like breathing rate, without conscious intervention. But a lot more than that is going on in your brain outside of your conscious awareness.

I should keep a list of my references because it's hard to search for obscure brain studies.

10 Weird Brain Disorders That Totally Mess With Your Perception of Reality

Symptomatic cerebral akinetopsia is defined as the conscious loss of visual motion perception due to extrastriate cortical lesions.

I don't know if this is an example but if it were, the person could react physically to the motion despite being blind to it. If something comes at them they'd get out of the way, but not consciously direct their muscles to move.

In one study a woman could not draw the angle of a slit. But she could directly put her hand correctly in it. I'll have to keep looking for the link. Her hand function did not need conscious thought to perform controlled movements.

7. Aphantasia: people with this condition can't picture things in their minds.
A 65-year-old man called MX suddenly lost the ability to summon up images of things in his mind after a coronary angioplasty surgery. On questionnaires, he reported not being able to visualise any images, despite the fact that he performed normally on standard tests of perception, visual imagery and visual memory.

After researchers reported this, more than 20 other people contacted them to say they had the same inability to picture things in their mind's eye. Though not yet a recognised neurological condition, scientists have eloquently dubbed the phenomenon 'aphantasia', from the Greek word for imagination.

The point of this is one's conscious self is believed to occur in a limited structure in the brain. It is not controlling everything we believe to be controlled by conscious thought.

Until we figure out how that structure works, all we can recreate with AI is a lot of memory and some complex algorithms that really amount to faux consciousness. You can create AI that can pass the Turing Test, but our conscious experience is not a function of a lot of memory and learned responses.

jjdebenedictis · Sep 19, 2018

OldHat63 said:
Oh, and Lizmonster... how 'bout bein' a dear, and not keeping me in suspense? What are the other movies that "trigger" ya?
Might be useful information down the road... keepin' an old fool like me from sayin' the wrong thing n' such.

Wow, that was gross and patronizing.

OldHat63 · Sep 19, 2018

Anybody else having an urge to find Ralph's side of the island and go have a long conversation with MaeZe?

'Cause I am.

O.H.

OldHat63 · Sep 19, 2018

jjdebenedictis said:
Wow, that was gross and patronizing.

Wasn't intended to be. Was intended as a joke.
But you're free to take it as you please. Provided you remember you're taking what isn't being offered.

O.H.

lizmonster · Sep 19, 2018

MaeZe said:
Take typing, most of us are pretty fast at that. Do you ever catch yourself typing a different word than is in your conscious thought? And I don't mean spellcheck, I mean catch yourself in the act. You think 'your' and your fingers have already typed 'you' and moved on to the next word. Just now as I typed the above, I consciously thought 'to' but my fingers typed 'the'. How often have your typing fingers gotten ahead of your thoughts?

After I learned to touch-type, I found there were a set of typos that I made habitually. And when I was writing longhand, I'd find myself misspelling things the same way. It was weird.

I'm not a biologist, but just based on my own experience (and my experience writing software), I don't believe the human brain is just a really, really complex data processor. Or maybe it is, but the number of variables is massive (hormones, hydration, age, usage/lack thereof, etc.), and we've yet to be able to reliably predict what causes the brain to be more or less effective at solving everyday tasks. Never mind disorders such as depression or ADHD, or why the same medications can be incredibly effective for some people and useless for others.

We're getting very good at tasking computers to do specific, targeted tasks more quickly and efficiently than humans can. That's nothing close to the spark of sentience except in crudest imitation.

(And to OH: The other films I hated don't touch on AI, and I don't want to derail the discussion. I know and respect a number of people who found Ex Machina profound and thought-provoking, so my objections really aren't germane.)

OldHat63 · Sep 19, 2018

lizmonster said:
(And to OH: The other films I hated don't touch on AI, and I don't want to derail the discussion. I know and respect a number of people who found Ex Machina profound and thought-provoking, so my objections really aren't germane.)

Thanks Liz. I really was just curious.
And I do hope you didn't find my manner of asking offensive. I apologize if you did.
It's just how I think and speak. Nothing more, nothing less.

By the way, I find myself also making the same typos, over and over again... It's almost like some sort of physical dyslexia. Weird.

O.H.

Introversion · Sep 19, 2018

OldHat63 said:
Wasn't intended to be. Was intended as a joke.
But you're free to take it as you please. Provided you remember you're taking what isn't being offered.

While we all look through our own lens, I also found it patronizing. "Be a dear" is not respectful to your fellow authors.

If one person tells you a thing, maybe it's nothing. If two, maybe it's worth listening to.

MaeZe · Sep 19, 2018

Disembodied fingers have a will of their own. Do you ever wonder how your fingers can find the keys without you looking?

They are habituated, but something different is controlling them than conscious visual cues.

OldHat63 · Sep 19, 2018

Introversion said:
While we all look through our own lens, I also found it patronizing. "Be a dear" is not respectful to your fellow authors.

If one person tells you a thing, maybe it's nothing. If two, maybe it's worth listening to.

Kind'a waiting for the person I SAID IT TO to comment, since I have spoken to her several times before, in an earlier thread and elsewhere... and have also already apologized to her if any unintended insult or offense was found, since it was mentioned..
Add to that the fact I'm right here, and easily reached by her or anyone else that wishes to speak on that particular subject... :Shrug:

Or is that somehow not good enough?
And if that is in fact the case... does that say more about me... or you?

Anyway, I really am convinced that Bronn of the Blackwater was right...

One way or the other thpugh, I am who and what I am, and am sure as hell not very likely to change now. So if anything I write offends you... how 'bout you just not read it?

So, enough of this bullshit. If you have a bone to pick with me, do it elsewhere. You know how to reach me, and I assure you I'm all "ears"... at least if you have good sense, anyway.

O.D.

lizmonster · Sep 19, 2018

MaeZe said:
Disembodied fingers have a will of their own. Do you ever wonder how your fingers can find the keys without you looking?

They are habituated, but something different is controlling them than conscious visual cues.

I used to notice this about playing piano. There were pieces I could play after years of not playing - but only if I deliberately didn't think about where my fingers were supposed to go.

Come to think of it, I sometimes write that way.

AW Admin · Sep 19, 2018

OldHat63 said:
So, enough of this bullshit. If you have a bone to pick with me, do it elsewhere. You know how to reach me, and I assure you I'm all "ears"... at least if you have good sense, anyway.

O.D.

I have a better idea. You can go sit in the corner.

amergina · Sep 19, 2018

And here I was gonna lock the thread and deal with it in the morning, but AW Admin beat me to it.

MaeZe · Sep 19, 2018

lizmonster said:
I used to notice this about playing piano. There were pieces I could play after years of not playing - but only if I deliberately didn't think about where my fingers were supposed to go.

Come to think of it, I sometimes write that way.

I've been doing it lately, seeing if my fingers can find the keys without me looking. I just got ~95% of that sentence.

But it's different from the 'hands on the home keys' I tried in summer school typing class that my dad made me go to after I got in trouble. My fingers can seek out the keys from any position over them.

MaeZe · Sep 19, 2018

For anyone interested in exploring why algorithms cannot explain consciousness, this article, while more technical than I prefer, nonetheless reviews how our thinking about consciousness has changed over the centuries.

Brain. Conscious and Unconscious Mechanisms of Cognition, Emotions, and Language

Abstract: Conscious and unconscious brain mechanisms, including cognition, emotions and language are considered in this review. The fundamental mechanisms of cognition include interactions between bottom-up and top-down signals. The modeling of these interactions since the 1960s is briefly reviewed, analyzing the ubiquitous difficulty: incomputable combinatorial complexity (CC). Fundamental reasons for CC are related to the Gödel’s difficulties of logic, a most fundamental mathematical result of the 20th century. Many scientists still “believed” in logic because, as the review discusses, logic is related to consciousness; non-logical processes in the brain are unconscious. CC difficulty is overcome in the brain by processes “from vague-unconscious to crisp-conscious” (representations, plans, models, concepts). These processes are modeled by dynamic logic, evolving from vague and unconscious representations toward crisp and conscious thoughts. We discuss experimental proofs and relate dynamic logic to simulators of the perceptual symbol system. “From vague to crisp” explains interactions between cognition and language. Language is mostly conscious, whereas cognition is only rarely so; this clarifies much about the mind that might seem mysterious. All of the above involve emotions of a special kind, aesthetic emotions related to knowledge and to cognitive dissonances. Cognition-language-emotional mechanisms operate throughout the hierarchy of the mind and create all higher mental abilities. The review discusses cognitive functions of the beautiful, sublime, music.

It's still not the research lit I'm looking for.

MaeZe · Sep 19, 2018

Here's something closer to what I've been looking for, hooray for Science Blogs.

Unconscious brain activity shapes our decisions

Our brains are shaping our decisions long before we become consciously aware of them. That’s the conclusion of a remarkable new study which shows that patterns of activity in certain parts of our brain can predict the outcome of a decision seconds before we’re even aware that we’re making one. ...

On average, the volunteers took about 22 seconds to press the button and felt that they consciously decided to do so about a second or less before they made the movement. But the fMRI data told a much different story. Two parts of the brain – the frontopolar cortex and the precuneus -showed activity that predicted the choices that the volunteers made and in the frontopolar cortex, this activity happened a whopping 7 seconds before the subjects were consciously aware of their decisions.

These astonishing results suggest that by the time we become consciously aware of a decision to move, our choices have already been influenced for several seconds by the actions of the frontopolar cortex.

I don't agree that because a process is occurring at the subconscious level it means we aren't freely choosing. I think it means we don't yet understand the intricacies of our brains and our decisions.

frimble3 · Sep 19, 2018

lizmonster said:
I used to notice this about playing piano. There were pieces I could play after years of not playing - but only if I deliberately didn't think about where my fingers were supposed to go.

Come to think of it, I sometimes write that way.

Crochet is that way for me, and sometimes, work.
Sometimes I can't think of the right move - if I stop thinking and just let my fingers decide, they make the right choice. I suspect that, like music, it wouldn't work for something new, but if the fingers 'know' the correct choice from past training, they simply follow the pattern.

*My work is traditional 'paper-shuffling', put forms in the right one of a dozen slots. Once I learned them, my fingers could do the walking, if the brain got sidetracked.

An examination of A.I. and E.I. using Ex Machina and other movies as a reference.

OldHat63

onesecondglance

pretending to be awake

lizmonster

Possibly A Mermaid Queen

bwebs

Harlequin

Eat books, not brains!

lizmonster

Possibly A Mermaid Queen

bwebs

Harlequin

Eat books, not brains!

OldHat63

MaeZe

jjdebenedictis

is watching you via her avatar

OldHat63

OldHat63

lizmonster

Possibly A Mermaid Queen

OldHat63

Introversion

Pie aren't squared, pie are round!

MaeZe

OldHat63

lizmonster

Possibly A Mermaid Queen

AW Admin

amergina

Pittsburgh Strong

MaeZe

MaeZe

MaeZe

frimble3

Heckuva good sport