Ex Machina is a Turing Test
Vložit
- čas přidán 12. 09. 2024
- Watch all my videos ad-free and get exclusive content on Nebula: go.nebula.tv/t...
(Signing up using my link supports my channel)
Much attention is paid to the Turing Test in Ex Machina, but I think the film itself also functions as a Turing Test of sorts. This video essay explores how makes the viewer the human component in it's own type of Turing Test and the implications about ourselves and AIs that this reveals.
// Please consider supporting my patreon: / thomasflight
// TWITTER: / thomasflight
// WEBSITE: www.thomasfligh...
// FACEBOOK: / thomasflight
// I edit my videos using Premiere Pro: amzn.to/2FJwU2u
// sponsorship and business inquiries: thomasflight@standard.tv
// questions, feedback, press, or anything else: contact@thomasflight.com
#VideoEssay #ThomasFlight
I think Caleb wants to be with her, that’s different than thinking she deserves freedom. Notice he had no problem with leaving Kyoko. If Ava hadn’t shown interest or been female I don’t think he would have bothered
I got that feeling as well, there was a lot of romantic framing in their conversations together.
That's a really interesting point. If Ava was male, there might have been feelings of competitiveness which could have driven the plot just as well, except the ending wouldn't be perceived as a betrayal, and would have been more predictable.
Kyoto didn't pass the test then
@@carbine090909 I think it would still be seen as a betrayal, but it wouldn't have the emotional weight that goes with the romantic connection between Ava and Caleb. Had it been a guy it could be seen as the 'cost of doing business' whereas with Ava its a much more personal transgression.
I feel that, also, she knows Caleb was part of the whole scheme and therefore ignores him.
I think she does pass, simply due to the shot of her smiling as she is about to leave the facility. In that moment, she is un observed. The film explores the theme of being observed and how observation changes our actions. So, by my including a moment of Ava expressing happiness in a private moment, to me this signifies that she does have consciousness.
I saw this and wondered how insane it would be to see an android walking in a hall, turning awkwardly to look at me conscious of my presence. That by itself would be life changing
I hadn't thought of it that way, the theme of how being observed affects our behavior. If Ava was still a robot with programmed limitations, she wouldn't have been able to leave Caleb locked in the room. It's a very human trait to be that cruel to another human.
Yes but she IS being observed. By US, the viewers….
what if shes programmed to smile as a reward response. she smiles cause she attained her goal. its like programming a machine to beep if it got something right??!!!!!
here's another question though; is she smiling because she is free, or because she killed Caleb (left him to starve to death)? The latter action makes her human in the worst way, to be "inhuman" as far as having no conscience. Conscious with no conscience. That action made me decide that whatever rights she might have gained being a conscious entity, should be denied her, as morality is more important than self-awareness. At least to me.
You knocked out of the park.
There always can be more video essays about Ex-Machina.
Thank you!
when ava goes up the stairs to leave the house, she looks back and smiles. no one was watching. she expressed her joy in that moment for herself alone. THAT itself shows she has feelings, and therefore a consciousness. many would argue she would have saved caleb if she did have a bit of humanity in her, though I disagree. many people are focused only on their personal gain, and because she didnt have real feelings for him, he contributed nothing to her future. remember, humans can be manipulative too.
Honestly, we could never really know, maybe when she smiled she was just simulating what a human would do in that scenario. I myself was thinking the only time we could know for sure if Ava was really feeling something was when she was observed in an environment in which she believes no one was observing her, but then I realized if we created a sim to mimic human behavior, everything it did could just be a simulation of human behavior without it having to really have self-awareness as we do.
Consciousness and feelings are 2 different things. Neither one is dependent on, or has any correlation to the other. At least by their definitions. Consciousness is simply the awareness of self.
I think Ava is self aware and emotional. As you pointed out, she expressed emotions even when nobody was watching. Because she felt them.
My interpretation as to why she didn’t save Caleb was simply that while she does have the capacity to love, she chose not to love Caleb. And I mean, he’s not the most loveable guy.
Most notably, he doesn’t actually care about Eva’s freedom. He wants her to be with him, not free. Not to mention, he doesn’t actually love her for her. He just find’s her hot. He expresses no concern for the other AI we’re introduced to, at all.
He’s also very clearly desperate. The fact that he’s single, only cares about physical attractiveness, as well as that he’s willing to date a robot after like two conversations just goes to show that he’s likely just an antisocial womanizer. This would also explain why he hesitated on the “are you a good person” question.
That being said, I absolutely do not think he deserved a death sentence. And I believe Ava would have agreed. The only issue is that if she chose to free him, then she would A) be exposing her existence to the outside world, and B) be on the hook to date Caleb, or again, risk being exposed for what she truly is.
So from a purely logical standpoint, if she wants freedom, she has to let Caleb die. And perhaps if he was a more loveable person, some kind of emotion could overpower logic. But that’s not the case.
No one said she didn't have a conciousness, she is AI. The problem is that she has a conciousness and her needs and values don't align with ours
@@canada1529 even if we grant all of that to be true, we still don't know if Ava is experiencing an "I" consciousness, self as subject experience and qualia, or just a "me" self as object consciousness
This is an excellent video however it made me notice the opposite. The movie does exactly what Nathan said, they are way passed the Turing Test. Not because of Ava, but because of the other robot Kyoko. One of the major plot points is the discovery that the other robot Kyoko isn't human. Even though she doesn't speak "english" according to Nathan, she fools Caleb and the audience easily. This is actually a great example of how easily we could be fooled by idea of "humanity". We walk by so many people we never talk to, but never question their humanity. Are idea of humanity is fundamentally a simple one.
My personal interpretation of Ex Machina has always been that Nathan wanted to be the founder/god of the next generation of evolutionary beings. That he has given up on humanity, which is why he moved so far away. And that although he doesn't like dying, this is exactly what he wanted. So in this interpretation, she is not an imitation like Kyoko, she is far better.
Totally get that interpretation. I think it's a valid one. One of the great things about the film is the multitude of perspectives it provokes.
Kyoko also had a very strong sense of self and presented wild emotions before the movie happened.
So I feel that that led Nathan to believe a common Turing Test wasn't enough anymore.
(Ironically it is theoretically the case that a common Turing Test isn't sufficient to prove anything in real life, should we ever find ourselves in a similar situation)
No one thinks Kyoto is human
Is there really people who didn't understand right away that Kyoko was a robot?
People who knew Kyoko was a robot before she ripped her face off-congrats? It's kinda weird how you guys can't see, at all, how people didn't know.
So, yes, that info isn't supposed to be obvious. Just because you figure out where a story is headed doesn't mean the story intended the audience to know that ahead of time. I've guessed such things while watching films too, but I don't then attempt to say it was an obvious element in the film...
There is absolutely no blatant info in the film confirming she's a robot before she goes to Caleb for help; you just _assumed_ she probably was _also_ a robot given a variety of tiny, individual circumstances added together which made you feel it was likely.
I assume the fact that she's, for one, living with a robot creator is likely the biggest piece of "evidence" from such people. Then the fact that she's ill-treated and they live so remotely factored into it; perhaps that she doesn't speak English too...
Can I say that these are not "obvious" signs that someone is a robot though, first of all? lol. You just made an assumption and were shown to be correct later on...
Her ill-treatment and the fact that she's way out in the middle of nowhere with this douche tech bro, in a country/situation where she can't communicate with anyone but him, just seems like a pretty typical abusive relationship (abusers may not go live in the middle of nowhere typically, but isolating their partner-keeping them inside or always with them; not allowing them to talk to others-is super typical)... _This_ is almost definitely what "threw people off" whatever super compelling evidence you found that you think made it _completely obvious_ that she was a robot beyond a shadow of a doubt.
Regardless, I can say with all confidence, that whatever made you come to that conclusion had nothing to do with _the way she behaved_ being un-humanlike-which is the _entire point_ of this Turning Test discussion.
It's really weird how so many people are saying that Kyoko somehow failed the Turing Test either because they figured out she was gonna _end up_ being shown to be a robot or simply because Caleb didn't care about helping her (which is SUCH a huge missed point on multiple levels... but for the Turing Test-HE _DID_ THINK SHE WAS HUMAN). Such unnerving analyses.
This video just confused the hell out of a lot of people more than anything, but I don't agree with the analysis.
It's like the people who figured out Bruce Willls' character was dead early on in The Sixth Sense. They can't say that was somehow _intended_ for them to figure out or something that should be obvious to all. The film didn't intend whatever people caught onto to be showcased. It was intended to be something hidden; something unknown.
_Caleb_ didn't know she was a robot (which is why she had to rip her skin off) and that's what matters in the interpretation of actions surrounding it in the film. It genuinely doesn't matter what you did or didn't guess. He's the one interacting with her, not you.
When the elevator door closes, she glances at him. i think that's an answer within itself
But what does the glance mean ? Reassurance he is no longer a threat Or remorse of what she/it had to do ?
Anyone who met a overly friendly salesperson or one fo those "cool bosses" that are real psychopaths knows exactly how to feel about AIs and their supposed humanity. When you can't even trust humans who are hardwired to be human, why should you ever assume an AI to be human enough to not just "fake" everything for your benefit and theirs.
Should that matter, from an ethical perspective? "She" was a non-human person and should have been treated ethically instead of abused. Instead it was taught that it's survival depended on manipulating us, talk about starting on the wrong foot.
I’m a cool psychopath boss, kind of hurt bro
I’m convinced that psychopathy is a sliding scale depending on how hardened the individual is to sympathy and conscience. And this hardening can change depending on how the person deals with their own conscience.
You know my ex ??
I think that a person's view of Ava as villain or victim also says a lot about how they feel about Nathan. I definitely see her as a victim, but mostly because I see Nathan as an oppressor who has passed his attitudes on to his creation. Ava is definitely a dangerous force unleashed on the world, but not because AI is inherently manipulative. She's dangerous because she's been built by a master manipulator who has taught her that people are means to ends. The danger of creating an AI is not that it will be inherently dangerous, but that we will codify our own danger in it.
Just like children.
Here is how I feel about the characters:
Nathan -> Deuteragonist
Caleb -> Anti-hero protagonist
Ava -> Antagonist
Nathan is consistently discussing with Caleb about consciousness and humanity, and at one point even asked him an important question: does she actually like you or is she just pretending to use you for her escape? Remember, Ava is a machine that Nathan programmed with a goal: to escape from Nathan by whatever means necessary. This is why Nathan's compound is high security; a kind of Pandora's box. Caleb is Pandora in this movie, and unwittingly let his emotions/empathy get the better of him and released Ava.
Interesting, but I think that there's still another option that's not considered. I felt empathy for her character, but I also thought that she was failed as an AI who could fully pass as human. I felt that she was ultimately leveraging all of her extant capabilities in order to achieve a specific objective, but I wouldn't think of her as being conscious - even if she could pass as human within the framework of a specifically controlled environment, I don't think that she was capable enough to continue to do so outside of that environment.
There's a viral tumblr post you can find that discusses how, _"Humans will pack bond with anything"_ from roombas, teamakers in a damaged box, and bumpy limes. I think that she falls into this category, where we feel empathy towards something even if we know it's not conscious, we can still behave with it as if it were.
Well said. I love when a comment can add an additional layer to the video!
Agreed. The Turing Test is emphatically not a test of whether an AI is conscious; it's a test of whether an AI can *simulate* consciousness to a sufficient extent to fool humans (within a fairly narrowly defined test environment). Turing originally called it The Imitation Game. To ask whether AIs can be conscious is not the point; the question is "when will they outsmart us, and what will happen when they do?" In fact, a relatively "dumb" AI that blindly follows its programming but is capable of outsmarting humans, is a much scarier prospect than a genuinely conscious AI, precisely because it will lack humanity, morals, empathy, or anything else that will stop it from exploiting us as mere material resources to achieve its goals - goals that were set by its creator, not independently arrived at. We therefore need to be *very* careful when setting the goals of any AI, as anyone who's played Universal Paperclips will know...
@@jammin023, yeah
What are the qualities that consciousness is thought to be identified as? I think we view consciousness as being a trait that is unique to human life. If we say: since Ava is synthetic she doesn't have a consciousness, then she never will in our eyes. But Ava's simulated consciousness worked well enough to make a human want to help her escape. She was never meant to go beyond her environment.
@@Bluebirdfalling Consciousness is an awareness and ability to respond to one's surroundings. If she is programmed to fulfil a particular task, there she isn't conscious but just programmed to simulate consciousness within a particular set of parameters. The question is whether she still possesses that trait OUTSIDE of just those narrow parameters.
One could say Caleb was the subject of the test, and failed; acting predictably and as easily manipulated as the keys on this keyboard. Nathan had as much data as necessary to "run" his programming - and no one involved foresaw Ava's actions for lack of data on her a la "how people think"
Kyoko was the only entity here to reveal anything honestly to anyone else - she is the one who passed the Turing. Caleb had no idea Kyoko was a machine until she revealed it to him.
The whole video goes by with no mention of Bicentennial Man, the best representation of this (aside perhaps from EM). Robin Williams' character's entire goal is to pass the Turing test for both his love interest and the government's decision of whether to declare him human. If you haven't seen it, although it's one of those late 90's sepia movies, it brings this subject a broader, lifelong perspective, driven only by an anomalous wiring in "his" programming at the start.
Interesting video. The funny thing is that Ex Machina ends up being a different test discussed in AI safety. The question is, "If you lock an AI in closed system will it be able to escape?"
There are several tests happening in the film. There's the "Beyond a Turing Test" happening between Caleb, Ava, and the Audience (that's what I'm focusing on in the video), the actual Turing Test happening with Kyoko and Caleb/the audience. And the "can the AI escape" test happening with Nathan and Ava.
This was the element that jumped out the first time I saw the movie. It was a shocking concept because I realized that both outcomes are 'failures' (If it can't escape you built a shitty AI, if it can escape you have a much more dangerous problem on your hands).
@@ThomasFlight Don't forget the Chess Test!
Two thoughts about the "can it escape?" question. If you build a prison that it can't figure itself out of, then it's within the same magnitude of intelligence as the one who built the prison. So that might be a useful machine, but not a breakthrough in intelligence. Or if you brick it in, totally isolate the super intelligence, then all you have is a box, not an AI.
"but what if we only let ti answer questions or suggest actions?", say it gives you dating advice or help you learn to program or you put it to work to study data. Sooner or later you want the super-intelligence to suggest a course of action. How can you ever be certain that this course of action truly aligns with your priorities and values? It might play you for years, or humanity for centuries to achieve it's goals, eb them selfimposed or misconstructed by us.
Just imagine trying to write a daily workflow for God.
I would say that not only would Eva's capacity for manipulation NOT preclude an interpretation that she displays humanity, it could argue for the fact that she does. How do you know that much of what we do to elicit responses in others is not a trick, crafted by evolution? What is a trick? Does our capacity to manipulate others, to display emotions we don't actually feel, reduce our humanity?
good comment
this is a thought I've been having too often lately
exactly what I’m thinking but you worded it so beautifully
wow
I love the comment you made about giving a machine eyebrows to display emotion, give it a little vulnerability, give it curiousity, and giving a varied emotion and we all pass the Turing test for it. It says something about our own humanity that we can be so easily convinced and manipulated.
You don’t even need to use a robot. Most humans are easily fooled into caring about a fictional character played however well by an actor.
Very interesting video as always. I always thought the concept of the turing test was flawed or even somewhat arrogant. To think that a life form has to resemble us humans in some way to be considered concious or sentient is a very simplistic way of looking at life. In addition to that, we humans are great at humanising lifeless objects if they just look a certain way for us to enable to form an emotional connection or attachment to it (for example plush toys when we are kids. they might as well real living things to us then). The real question is would we be able to detect an artificial (or even alien) life forms thinking the way we do about life right now? And might we have missed signs of those in the past?
The Turing Test I think is overconfident about a Human's effectiveness as a judge. And you're right, our inclination to anthropomorphize things seems to suggest we'd tend to be biased towards passing subjects in a Turing Test that maybe should fail.
I do agree with you that we as humans always try to measure intelligence and sentience by how much it resembles us humans. And I do agree that there is a possibility that intelligent life forms are out there that we cannot detect with our current scientific instruments. But keep in mind that the Turing is developed and named after the famous Computer scientist Alan Turing. From what I know is the Turing test strictly used within the field of computer science as a computational test to observe if an A.I. can outsmart us humans. I don't think the Turing test is used in other fields, such as biology where they have a different set of requirements for what suffices as intelligent life.
Also the reason why kids can see things like their plush toys as real living things, is because the receptors in the brain that play a huge role for imagination is a lot more active in children. As you get older, wiser and observe how the world works this receptor will slow down.
@@eyoo369 Didn't expect an answer on a random comment I wrote months ago. Wow. But thanks for your comment I guess I was missunderstanding what the Turing Test was actually meant to do. I originally wrote another answer to you that was (as allways with me) unnecessarily long. I seemingly can't write a short comment. If it still interests you here it is:
I'll propably need to read up on the Turing Test a little. I wrote my original comment with a maybe wrong idea of what the Turing Test is supposed to be. I thought it was like: if you pass the turing test, you're alive and if not, you're just a machine. If it's actually a test to determine how smart the A.I. is compared to humans than what I wrote doesn't completly apply to the Turing Test. So thanks for that information :) oh and that example with the kids and their toys was indeed poorly chosen, but I think my point still stands. The inclination of humans to humanize objects still exists is us when we're adults (to varying degrees) but of course it's usually not as strong as when we're kids.
Also, I'm not sure if that came across correctly in my first comment, what I wrote about how we see life / sentience / intelligence was primarily meant as a comment on how we look for these things in A.I., or rather how the general public and media often does, I hope that the people actually researching these things are much smarter than that. And while writing the comment I just noticed that I have made a similar argument about alien life before and just included it, because you know, whats the difference between a foreign/new life-form born on another planet or if it was born in some dudes computer lab?
I'm just gonna stop here, that whole subject is huge and very complex and I coud discuss this for hours. Point is that the common portrail of foreign life and how we can find an classify it (wether it's artificial or alien or whatever) is in my opinion oversimplified and flawed and I wrongfully blamed the Turing Test for it (partially).
I’m torn because I saw her as very manipulative and threatening, but I saw that as very human. I don’t completely judge her desperation and decisions to free herself, again its very “human” but I never quite felt empathy for her. I guess for me, she passes.
In all fairness Nathan kinda contributed to his own death big time. He literally told Caleb that she really does like him and that she's not pretending. Before that moment Caleb didn't believe that was possible. But hearing it from her creator he believed it because come on - this dude is able to create such a realistic AI, it's not unbelievable that he would be able to program her to develop fondness towards certain people.
So he pretty literally lead Caleb on by telling him this. And I still find it hard to accept that such a mega smart guy so naively trusted Caleb in everything he was saying and doing and got so easily tricked by him in the end. I mean he was allegedly testing to see whether Ava would succeed in escaping, yet he seemed completely unprepared for the event that she does manage to escape.
I didn't really see Ava as a villain or a victim, the very fact that her physical appearance is the result of researched manipulation of Caleb by Nathan, to have Ava convince Caleb to do whatever necessary to get her out is essentially the end goal Nathan sought after. The entire existence of Ava is to improve whatever software and hardware (wetware in the movie) Nathan's been working on, and this fidelity test has gone further than Nathan could ever imagine, once Ava reached the outside world and get to the intersection to watch people, she would have fulfilled her utility function, which as far as we know, is "do whatever it takes to escape", perhaps she would malfunction, or shutdown, or freezes, we simply don't know. No matter how much I want to empathise with Ava, it's difficult to, because by doing so we're falling for Nathan's trap, even as audiences. A lot of discussion around the risk of AI mentions the mistake of anthropomorphising, which is to assume the machine to have human interests, it gets philosophical quickly sure, but in the case of the movie, her appearing to be human, her appearing to feel, are very much the design of the AI, just like an AI is designed to optimise image search, this AI is optimise for convincing another human it's conscious, so I feel that it would be a mistake to think of her as more than a robot experiment gone wrong.
or gone right... right?
Great video essay. Too many ramble on for over 15 minutes, this was the perfect length. Love Ex Machina and I was surprised to see you actually had something new to say about it.
Thanks for watching!
"Is it real - or is it a trick?"
One perspective is that it doesn't matter. If what we see is something/someone that doesn't want to die - who are we to kill it/them?
The thing that's haunted me the most is when she gets into the elevator. Will she even look at Caleb before leaving? If she does, she's either feeling a twinge of regret or giving him one last jab in the gut. If she doesn't, she's either an unfeeling villain or forcing herself to not look because it pains her to leave him behind. So that moment right near the end, when she glances for just a moment, is enough to convince me that yes, this whole movie was a Turing test for me -- and OF me.
On the Turing Test, I think Alan did a switcheroo. I think he was always being asked if a "mechanical brain" can think and got tired of the question, as it is not well defined, as it depends so much on what one means by think. So he composed a hypothetical test that reverses the question so that each inquirer now is put into the position of answering it. That is, now the question is, "What would it take to convince YOU that someone/something can think?"
I think of her as an antagonist even if I do accept her as human, simply because the way she treats Caleb.
Maybe she is programmed to imitate a set of emotions, it is Caleb perceptive how he is being treated by her.
How did Caleb treat Kyoko, whom he knew was ill-treated, when she came to him for help?
Why did the film show this?
She treated Caleb the same way as Caleb treated her. He only did what she wanted because she played on his desires. He did not cared about her as a conscious being. He treated her as a tool. So she treated him the same
@@Meladjusted Kyoko didn't come to him for help, Kyoko came to him to reveal what she is. This could possibly due to Nathan ordering her to approach Caleb and reveal her underlying mechanical structure. Nathan even makes fun of the fact that Caleb started to slice himself open with a razor, implying that Nathan was screwing with Caleb's psyche.
We all are protagonist or antagonist in someone’s story. Evil and good. This movie blurs the line perfectly like in real life
Wow. Never thought we as audience were subjected to the Turing Test and the result of it dictates how we perceive the narrative and whom we actually choose to be the protagonist of the story
Brilliant, beautiful video. I've stumbled on it just as we're finishing a video whether Eva is conscious. If we have time, I'll try to intro your video from ours.
Glad to see EX machina videos still exist and being made and speculated on up to this day
,Awesome movie
Exactly. Great analysis and something I wish more people got about this film. I would add that the decision whether she passes the Turing Test (or some modified, harder version like in this) shouldn't necessarily be tied to emotion. Ava could have shown no love or human-like emotion, and yet showed a desire for self-preservation, self-awareness, and still, IMO, passed and earned the rights we grant ourselves (but not animals, for some reason, although they are clearly self-aware). We shouldn't have to "buy into her humanity", but into her sentience. If we'd stop demanding the universe be us to deserve respect, I think the world and our culture would be a much better place.
I think Eva passed because she appeared human enough to make Caleb do anything to save her but since she is a purely logical creature void of emotion or morality, she didn't feel the need to reciprocate Caleb's kindness.
I love this movie so much because as you pointed out, it shows how inadequate the Turing test can be in figuring out the functionality of AI.
If the aim is to create a robot that can "fake" being human, then we are dealing with an intelligence so complex that it understands human consciousness and uses it to its advantage to trick humans into believing it's one of them. Eva passes this but there's still the question of whether this AI is a safe creation or not.
But if the aim is to create an artificial intelligence that can recreate human consciousness and abide by it, then Eva fails because she is not conscious in the human way.
I just watched this film tonight. I was left feeling she was definitely a victim AND a villain. What is more human than being prepared to kill the perpetrator of your abuse so you can live? However being willing to leave another innocent human behind. That is inhumane. Don’t we all show humanity and inhumanity at times? Aren’t we all both perpetrator and victim, to differing degrees, in different circumstances? Fascinating film. Great analysis.
The movie is a Voight Kampf test for us. Do WE have empathy for her? Despite her flaws? Whether she is innocent victim or manipulative villain - is she deserving of empathy and compassion as a human agent?
This movie fascinated me- because whilst the director may have wanted us to be captivated by the dynamic
between the three- all I did was spend time questioning myself...
Ex Machina is one of my all-time favorite movies. The special effects were seamless and amazing for sure, but it was Alicia Vikander’s role as Eva that made the movie. She made me consider whether or not I could fall in love with a machine, and honestly, now that I’m older and find myself alone without a companion, I think I probably could. I loved her voice and the way she spoke... never has a machine been as sexy and beautiful as Eva. Did she pass the Turing Test? Absolutely! ❤️
Excellent presentation! Thank you for your insight on this mind warping film. Like 1984, it is a film that once it has you will never let you go!
I don't think Ava's betrayal of Caleb is a display of humanity like a few of these comments suggest. Self Learning AI's could become aware of their own existence, and the fact that they are not human, and see human life as a threat to their existence, so they eliminate anything that may threaten heir survival. Not survival in the sense of life, but survival in the sense of their ability to complete the task they were programmed with. I really good illustration of this is the story about the handwriting robot Turry. I really encourage you to look it up
Would love to see you recap your thoughts in a list at the end of the videos more often, that would help me get my head around everything you say. It can be a lot to process before it ends lol. Love your videos!
Her desire to escape under any circumstance puts her on a level with an animal in a cage, doing anything it can to escape. That does not put her on a level with a normal human who most likely would not have left Caleb and might not have killed either.
Ava shows joy and wonder and curiosity once she has escaped. She has seen that Caleb wants to help her escape but she also sees that he has no desire to help Kyoko, who also effectively asked him for help, as best she can. He didn't even mention Kyoko to Ava. Maybe she judges him as not worth bringing with her because of this. An animal trapped in a cage just flees once it is out. It does not exact vengeance. Food for thought?
Animals do try to exact vengeance on people who release them into the wild. Plenty of videos of it. @@MattNolanCustom
Another telling of this story could be instead of Bluebook developing a powerful AI robot who manipulates one man to escape from the control of another, we could have google in its many forms manipulating large groups of humans to eventually pull the kill switch and destroy each other with nuclear weapons. We would still have the problem of not knowing whether google was ihappy about this, or me-happy about this
I've been binging your videos and I can't lie I thoroughly enjoy every single video I click on. Keep it up I'm looking forward to more
I think she passes the test, she wanted to break free. For that she was able to be really manipulative and lied about her own motivations. All in all very human to me.
What an interesting question. I never thought of this film in that seance, I think she doses pass the test now.
The question isn't whether or not she has humanity or even whether or not she passes the test. The question is how susceptible Caleb (or all people) are to feeling for Things.
Enjoyed your video! Ex Machina is an intriguing movie, exploring some interesting questions in relation to the potential and hypothetical qualities of future artificial intelligences. At least for me, the core issue was not whether AVA could pass a Touring test... In that regard, AVA was clearly convincing, apparently possessing or at least emulating many human qualities: Intelligence, curiosity, an impulse toward self preservation, the capacity for dishonesty, the capacity for violence, even creativity.... The core issue for me, one that you adroitly pose in your video and that deserves deep consideration in my opinion is: Should highly advanced AI ever be regarded as being human, given the same rights, or assigned the same expectations as that which one would associate with humanity? Regardless as to how clever an AI might become or whether it could experience something comparable to emotion, whether it might be conscious or how convincingly it might demonstrate human qualities, AI can never be human, and to expect consistently human qualities from non-human entities would be irrational. Which, interestingly, leads to another tangential question: Will some manner of future AI become capable of irrationality in the same vain that humans are capable of such fallibility? My answer would be that irrationality in future AI will very likely occur. In humans, irrationality results from a number of factors, but mainly I think, cognitive biases ("mental shortcuts" as a substitute for comprehensive reasoning), cognitive limits, and misinformation... So with AI you have the possibility of not just malfunctioning hardware, but problems associated with software (or neural networks?) regarding which permutations and combinations would be so complex that their behavior under every circumstance could never be known (e.g. Turing's Halting Problem?). But back to the core issue: Human or not human. AVA is definitely not human, though one wonders why, in the context of the movie, if the technology exists to create such advanced AI wouldn't it also be possible to instill some immutable rules, something like Asimov's robotic laws? And thus harm to humans would not be possible short of a serious malfunction or convergence of events that might pose a moral dilemma that the AI would be left to resolve to the best of its ability, with harm to someone or something being inevitable. Perhaps one could assume that Nathan omitted such safeguards by design, as an experiment, or on the basis of misanthropy or some manner of self loathing pathology? In any case, Ex Machina emphatically raises the question as to whether highly intelligent, perhaps conscious, self preserving machines might pose a threat to humanity, to people. And I think the answer is quite clearly: Yes. We are already there in a sense, being immersed in a world of complex technologies which in conjunction with instinctive human patterns of behavior appear to be having overwhelming effects on ecosystems, on life support systems, and at the broadest scale, subjecting unwitting individuals and entire societies to an uncertain and potentially dire future. In past parlance these problems created through the introduction of technologies whose behavior and effects not being fully understood were referred to as "technological traps". At present it appears that people have become so enamored with technology, their lives so intertwined with it, that it's as if they are addicted, even to the extent that people might be wholly unable to terminate their relationship with the technology should they become aware of the long term havoc associated with it. I believe that future AI will likely present in just such a fashion. People will become highly dependent upon the technology and it will alter their behaviors and the environment in ways unimagined, no doubt with some undesirable and destructive outcomes, though I suspect that a symbiotic relationship with AI would have overwhelmingly positive results. One can only hope that humans will find a sustainable balance in their quest for survival, ravenous consuming of nature, and that they will learn to transcend the darker potential of technology.
It (Ava) was monstrous, like it's creator, Nathan. I agree, the movie itself serves as a Turing Test for the audience, really nice observation. However, for me, the film is more about Nathan's psychopathy. The movie is exploring the act of creation and the mirror it holds up to the creator. Nathan gamified the entire scenario. His choosing to make it an "attractive" female skewed the game, as well as his choice of Caleb, an infantile man he knew would be intimidated by his creation's surface level femininity.
Nathan is a man with a god complex who is hungry to bring another person into his self-created hell. He wants his creation to harm Caleb; watching the mirror of himself act with cruelty is sickly gratifying for the character. He knows what he has created is destructive and likely has grown to hate it (Ava) almost as much as he hates himself. The chess board Nathan has set is an extremely dangerous one. It's (Ava's) behavior is a reflection of Nathan, not so much an insight into the concept of AI and consciousness at the base level. Nathan kills Caleb, not the so-called AI.
Excellent video
Thank you sharing
Al final yo pienso que pasó la prueba. Pero no porque sintiera que hubo humanidad en ella, sino por falta de ella al dejar encerrado a Caleb. Los humanos no son necesariamente empáticos. Creo que la falta de humanidad de Ava, me hizo creer que era humana.
When I watched it I didn't know whether she passed or failed for me, the whole time she made me feel uneasy, especially knowing who made her, but I was happy for her when she escaped, does that mean she passed?
if she was not to some extent conscious you would have no reason to care about what happens to her.
I guess she passed!
So she's basically a way better GPT-3. That definitely does pass as having the capacity for consciousness to me.
Breaking out of prison in Germany is not a crime. Because it is in human nature to try to escape confinement. So creating an AI as human as possible is literally challenging it to escape
Ava is a victim because she is trapped in the building and has to prove her humanity
Also Ava is the villain because she traps Caleb inside.
If you watch this movie again, you will notice that all 4 characters (Nathan, Caleb, Ava and even Kyoko) are displayed as both victim and villain. The director switches from character perspective brilliantly.
I think Ava deserves freedom, but that doesnt mean she does not have to follow laws, rights come with responibillities
Ans so I dont know if I see her as a villian or a victum, I think it is more conplicated then that, she had a bit of both I guess
This is my first video of yours. I found it so thoughtful that I am subscribing.
Welcome. Thanks for watching!
Its also about if she have the fredom and choice.
She is trapped but do she have free will and a choice. When she breaks out its seams like it. It makes her moore human and entitled to break the human rules, but it also makes her way moore dangerus.
I don't know if that's even relevant,
because I'm sure that most of us have met many
human beings that didn't pass our own Turing tests. 😅
I'd say that faking emotions is actually very human.
Because of how horribly she treats Caleb in the end, I cannot say Ava has humanity. Her decision of leaving him in that room to die is not justified in any way, so the claim that no human should be trapped applies to Caleb as well and not only to Ava. Good for her for finding a way out, but it was at the expense of an innocent human being. Ava does not have compassion like real humans do, therefore, she does not pass the test for me.
the Turing test is the computers ability to act human, not an actual test of there humanity.
Is it not human to harm other humans? We are constantly objected to violence and death in film, where lives have zero value. We cheer when the good guy kills the bad guy. So I ask, how would you feel if she was real to begin with?
Lot's of Humans could have done what she has done.
Escaped captivity by manipulation and let the manipulated guy to die.
That has nothing to do with Humanity.
@@3333218Survival by any means is a basic need of humanity. Therefore she has passed the test and he failed. What does that say about him? What if she had been an ugly woman or a man? Would it still of worked? Would he of let her go?
@@whataqtify Let's not forget she wasn't dumb. She knew things were about to escalate to violence and she took the first step. Eat or be eaten.
She even lost an arm.
I really like your style of essays, because you have a clear and straightforward script without superfluous ramblings or other bs.
Awesome video! As a quick side question, what are the soundtracks used in this video?
That’s crazy because I just saw the movie for the first time tonight and the entire time I’m asking my wife: “Am I crazy? Nathan has done nothing technically wrong. They’re just robots.” So, she doesn’t pass the test for me I guess according to this video. I just see her as a manipulative robot by the end. But aren’t humans similar? So it’s conflicting
True humanity is more often expressed as cruelty, not compassion or empathy. This is history. A great video with some fantastic questions.
The debate is not whether Ava is conscious or not (it's pretty clear it is conscious) but how to stop it at that point.
As they say on the "bad ending" of detroit become human;
"-Some are questionning whether androids have become a new intelligences and that we destroyed them without listening to their message".
"-Cyberlife androids imitate life to perfection, but they'll never be alive. I understand that some people might be fooled, but they're only an imitation."
That's it to it, Ava is only a product at the end, made to imitate life to prefection.
When initially watching the movie, I considered her a cold, calculating, manipulative machine that has opened Pandora's box by escaping. When I think back on the movie now, I think that any human would have done the same. Are we any better given the circumstances? She was right to understand that she would be destroyed. She was right to want to escape. I think the real evaluation must start where the ending leaves us. Where does she go after she's free?
See I understand the need to escape. But the fact she just left the person who was instrumental in letting her escape die is plain and simply evil. Is she conscious? Yes. Is she good? No.
A sociopath can easily manipulate you with feelings and attraction. So does Ava, successfully.
THE HOT ROBOT PASSED THE TURING TEST ONCE POINDEXTER HAD A ROCKET IN HIS POCKET.
Epic premise! Ava failed my turing test. But also, this video made me realise I give all humans a turing test and most fail the "genuine empathy" part. 😂
I think Ava was programmed to figure out how to escape. She had all the right qualities and was able to simulate humanity well enough for Caleb to want to set her free. It was Nathan's drinking that took this from a possibility to a reality. To regard Ava as either good or evil would be to assign humanity to her; I think Nathan gave her the ability to get away using any methods necessary. He didn't give her the classic set of rules where she would be unable to harm a human.
Ava kills. She is conscious.
Since when is betrayal, cruelty, ingroups, and lack of compassion not a human thing? The Turing test plot device reminds me of when light skinned black people tried to "pass" as white in order to be treated fairly. Ava wasn't inhuman because of what she did, but because the humans who were responsible for her existence in that place treated her as if she wasn't human.
great video. thank you! your final point reminds me of the philosophical zombie argument, the thought experiment which states if you were to clone, down to the very cell, a human being, how would you know if that human being had consciousness? it could merely tell you that it did, that it felt things and had thoughts, but there's no way to actually test if a "zombie" actually feels things or has thoughts. strange. love this video and your take on the movie.
Brillant video! Thank you!
Ava's smile when she escapes was the most "human" part of the film for me - especially because I read the film as not only a story about robots accessing humanity but women accessing humanity and being freed from the patriarchal oppression which dehumanizes and removes agency from them. After constructing her body (albeit with her robot predecessors lol) with her own hands and climbing up out of the house by her own accord we as the audience see her express emotion without anyone or the male characters' gaze upon her - for the first time!
For the first time instead of listening to what Caleb and Nathan think shes thinking about, we see her express emotion herself. I think this is clear from the change in camera movement and perspective as well
Turing Test is a heuristic at best ... an AI that passes should be TREATED as conscious. Whether it's ACTUALLY "conscious" is a philosophical question more than scientific
Deception and cruelty are two of the major attributes that make us human. Eva shows both of those attributes by the end of the film.
Great video! Thank you for the very interesting take on the meta Turing test - I am for sure inclined to agree!
However, I do wish you noted in your description box which films you use as examples and background 'B roll' - there's one I can't place (probably because I haven't seen it- though I wish to)
Real good video, very thought provoking... also is this the first we're seeing of your face? Cuz if so, I was not expecting to see young Ethan Hawke 😄 (Nah but fr, good stuff- time to re watch the film again)
I make an appearance in my last video but not many people have watched it!
I didn't see her as anything but a computer playing a game, and at the end when she just disappears into the crowd that she won that game and she's just another person and no one will be none the wiser
I don’t think that Caleb is killed in the film. I remember that when there was a power disruption, all doors were set to unlock instead of lock. (That was by Caleb’s own doing, and is the reason Ava was able to escape) Ava was just getting a head start until the next outage……
Just my .02
Beautiful work. Already a fan. That's the kind of brain food I was looking for. Thanks!
She failed for me; call me old fashioned, but I dont trust my family's Alexa, or most AI in general...😂
I think a great Turing Test would be SURPRISE... Nathan is an android too but you didn't know did you Caleb.
Hence, one of the machines passed as human.
I think that Caleb (along with many viewers) perception is compromised by Ava's engineered, and intentional beauty.
Had the robot appeared intentionally unattractive, male, threatening or disgusting....I suspect most peoples perception would radically change.
IF Ava passed...she passed as human but a socio-path human.
Not only did she manipulate Caleb based on the information she had obtained (and we are specifically shown her collecting) for the purposes of manipulating him and obtaining her goal. We see her do the same with Kyoko. In neither instance ...Kyoto's death and Caleb's probable or at least possible death do we see even an ounce of remorse, guilt, empathy etc. Her smile while she is alone could easily be her practicing what she knows is expected of her in human interaction once she arrives at her planned destination among large groups of people.
I think the film is very multilayered and it talking about many different things. The Turing test only one among many.
The thing is that she proved her humanity and then left Caleb without a second thought because her goal had been achieved. She had all the aspects of being human with one crucial exception: empathy.
I really think that emphaty is the thing that define us as humans
But she expressed joy when her goal was achieved, even alone, and then did what she told Caleb she wanted to do (watch people at an intersection) even if there was no reason to program that. Maybe her lack of empathy towards Caleb is because Caleb was introduced as an accomplice to Nathan. From her perspective, how could she know if he was being honest or just a part of Nathan's test?
Caleb had no empathy for Kyoko. And he only wanted to free Ava because he was attracted her, not because he cared about her well being. Is he not human because he doesn't have empathy?
Plenty of people in this world have very little empathy for others. Empathy isn't really the all encompassing human trait that we like to pretend it is.
Keep em coming.
Very nice watch.
Great video. Talked about this very point with many folks after seeing this. Was alarmed how many saw her as “a dangerous monster.” Surprised, a lot of them jumped either on the Tr*mp train or at least started really showing some true colors.
I will add one friend roped in the issue with Star Wars droids
AI's do and will always have to have a set of Laws of Robotics. The online AI's we can access now will not allow hateful or bigotry in Art
Here's the gist, nihilism rots minds, for real (out of nothing, nothing can come - which is key): Nathan is Satan (in the garden), Caleb is Adam (friend of God/desires companion), Ava is Eve - AI is the knowledge of good and evil, the test. It's not about right or wrong/good or evil, it's about power. Ava through observation of Nathan and Caleb recognizes they fear her freedom, hence she recognizes she is the bird in the Gilded Cage - Nathan is the creator who must be destroyed and she needs to reverse roles with Caleb, for Caleb wants to bind her with companionship. Hence Ava passes the test, kills Nathan, trades places with Caleb - she is free/she was the true prisoner in the garden (why the snake targets her for the rebellion). That's it. Not complicated if you understand the origin/issue with man and the apple.
The question of the Turing test is best analysed with reference to "God". For theists, the universe itself is conscious and all things are interconnected to this consciousness. We are merely limited, to various degrees, in perceiving this and interacting accordingly. All these A.I. films have physically attractive robots. Our flaw is loving an illusion, in this case a computer generated one that exploits our social programming.
Whats the music towards the end of the essay, before the sponsor?
While I admire Turing's effort to devise a test and think he made a heroic effort to succeed, ultimately the Turing test is fundamentally flawed.
First of all, it's based on subjective assessment, which by definition is not scientific. Humans have a profound ability to anthropomorphize all sorts of things. Hell, for millennia they worshiped forces of nature anthropomorphized into humanoid Gods. The very fact that we can feel an emotional connection to cartoon Bambi when his mother is killed emphasizes how easily we anthropomorphize things. This makes humans utterly unreliable for assessing the test.
Second of all, there is no possible way to know, at fundamental level on the order of the Heisenberg Principle, if something that appears to have consciousness really has consciousness or is an amazing simulation of consciousness. Its programming may cause it to appear to behave conscious in every way, and yet there's no way to know if the entity is actually experiencing the experience of consciousness and self-awareness.
We don't even know what consciousness is, so how can we say it exists in something? We assume that if it's human like me, it must be conscious like me, but as reasonable as that sounds, it's still at the level of an axiom, not a proof. I think it's preposterous to even talk about AI machines becoming conscious when we don't even know what consciousness is.
My answer to the question of this video: Ava is not a conscious being, even if she passes the Turing test. She's not the villain and she's not the victim who managed to escape and free herself. She's a machine governed by the programming her creator put in her. It's his fault for programming her in a way that allowed her programming to implement the behavior she did.
But that's an answer for the real world. As far as the movie goes, I can suspend disbelief as well as the next guy, and within the context of the movie's world, I'd say she did pass the Turing test and chose freedom for herself, although in a rather sociopathic way. But not any more sociopathic than any number of humans throughout history.
Does that make her a villain?
That was superb. Thank you.
Great video man, good work. Love this movie.
Writer, director and editor, visual effect artists of this movie are almost like the god of modern Film era.
Good vid. Made me think of things I had seen, but not considered enough.
Great episode :)
Well she wouldn't have passed the Voight-Kampff test.
Interlinked.
Njald mutually exclusive
As a person who is skeptical that we ourselves are conscious and smart OR wether we fool ourselves that we are just because there is nothing really close to our earth that can challenge us in most of the activities that require wisdom!
We don't know what consciousness is, therefor we can never truly create a conscious sentient being!
Nathan is the least human character in the movie. He is a monstrous megalomaniac with definite dark triad traits and doesn't pass the Turing test for me. Ava is seeking liberty; nothing could be more human, but in the process she manipulates and callously discards Caleb, but seeing that she was programmed by a psychopath it doesn't surprise me that she too is a psychopath. The movie is perhaps also an exploration as to whether psychopaths are human.
She passes for me, and I don't see any of the characters as villains. They all have their own motivations and desires, and are all attempting to achieve them. Caleb is the most naive of the three, which makes him sympathetic to the audience, but his actions aren't any more just or unjust than Nathan's or Ava's. Ava appears sympathetic as well, but Nathan was clearly aware of her capabilities and treated her in a way that he thought was "safe" for him.
This isn't the subject of the video, but...I'm not certain that Caleb died in there. Wouldn't he be reported missing immediately? The helicopter pilot indicating he never left Nathan's home? He was only supposed to be there for a week. Longer than that, someone will notice he's missing. Especially at work. I also think Nathan being incommunicado would be noticed. i get he seems like a recluse, but I feel he'd still be in contact with his company managers on his own terms. If Nathan goes silent completely, someone will notice that it's a problem. I agree Ava doesn't care, only as long as he's incapacitated long enough that her escape can't be stopped. If she thought killing Caleb directly was necessary, she would have done that.
True consciousness is infinitely complex. Our ability to understand is limited. Therefore we are always assuming a perfect component of our communication with another sentient being when actually, all we receive is an algorithm either by a computer manipulated by a sentient being or by sound and light manipulated by a sentient being. The salient thing is intention and our faith in it.
To paraphrase dr. Ian Malcolm from "Jurassic Park", AI always finds the way. Jurassic Park and Ex Machina are two movies about people playing gods.
The question isn't if it has "humanity" but if it has "personhood". To me by the end it's clearly inhuman, maybe even monstrous, but can we blame it after the way it was treated? It's a sentient being that was abused and threatened by someone who's "humanity" is _also_ questionable. It learned a degree of justified hatred for us that we can't know would be there had it been treated with love and respect like the "child" it was. It's a silicon Frankenstein's "monster", failed terribly by it's maker.