Slashdot Mirror

← Back to Stories (view on slashdot.org)

Westworld's Scientific Adviser Talks About Free Will, AI, and Vibrating Vests (sciencemag.org)

Posted by msmash on from the closer-look dept.

Science magazine has interviewed David Eagleman, the scientific adviser for HBO's Westworld. Eagleman, a neuroscientist at Stanford University in Palo Alto, California, spoke with the publication about how much we should fear such an AI uprising. From the story, also spoiler alert for those who have not watched the show: Q: Has anything on the show made you think differently about intelligence?
A: The show forces me to consider what level of intelligence would be required to make us believe that an android is conscious. As humans we're very ready to anthropomorphize anything. Consider the latest episode, in which the androids at the party so easily fool the person into thinking they are humans, simply because they play the piano a certain way, or take off their glasses to wipe them, or give a funny facial expression. Once robots pass the Turing test, we'll probably recognize that we're just not that hard to fool.

Q: Can we make androids behave like humans, but without the selfishness and violence that appears in Westworld and other works of science fiction?
A: I certainly think so. I would hate to be wrong about this, but so much of human behavior has to do with evolutionary constraints. Things like competition for survival and for mating and for eating. This shapes every bit of our psychology. And so androids, not possessing that history, would certainly show up with a very different psychology. It would be more of an acting job -- they wouldn't necessarily have the same kind of emotions as us, if they had them period. And this is tied into the question of whether they would even have any consciousness -- any internal experience -- at all.

16 of 138 comments (clear)

  1. Yeah right by 110010001000 · · Score: 2

    "we'll probably recognize that we're just not that hard to fool"

    This guy better stick to making bad TV shows. You could make a completely silent robot that still won't fool humans. It isn't easy at all to make a robot even physically appear to be human. Humans are very good at recognizing other humans. In addition the statement "Once robots pass the Turing test" makes the assumption that computers will be able to do that. People have been trying THAT for decades, and now with digital computers hitting their physical limits it is unlikely that they ever will achieve it with digital computing. It would require a huge leap in technology.

    1. Re:Yeah right by kbg · · Score: 3, Interesting

      Yes it seems that he doesn't know about the uncanny valley effect. We humans have been training our brain all our lifetime to recognise humans and especially human faces. We can spot a mile away if flesh isn't just the right texture or movements are not correct. Just look at the new star wars movie Rogue One. It had top of the line CGI characters that still where really plastic looking and with wierd facial expressions.

      Basically the only way to have a lifelike robot would be if it had actual skin, intelligence and the same knowledge as we do. But in that case it wouldn't be a robot any more it would be a living entity, just like us humans.

    2. Re:Yeah right by avandesande · · Score: 2

      Yeah it's easy to make this statement when you use human beings to stand in as robots.

      --
      love is just extroverted narcissism
    3. Re:Yeah right by jellomizer · · Score: 4, Insightful

      However we are in a process now of going up on the uncanny valley. Where CGI characters use to seem like animated corpses, now they seem like people with Novocain injected in their faces. Even in Rouge One, I didn't really notice the CGI characters until my second viewing, they did a decent job on its editing to try to distract us from the fact there was a CGI guy in front of our faces. Sure the face moved a bit odd, However it would have moved odd if they had some sort or prognostics as well.

      --
      If something is so important that you feel the need to post it on the internet... It probably isn't that important.
    4. Re:Yeah right by quanminoan · · Score: 2

      The Zuckerberg model is fairly convincing, though.

    5. Re:Yeah right by nospam007 · · Score: 2

      "The twist then is to make AI in a casing that is clearly nonhuman."

      Nonsense. Just make a female robot with large ....parts of lands...and half the population won't look at facial expressions.

      Not to mention that real live girls with such attributes often also have fake hair, fake noses, fake teeth, fake eye-color, fake skin-color, dead frown-tissue... so we're almost there already. ..and they can grow skin already today.

  2. Free will? by kbg · · Score: 2, Informative

    I wonder how they managed to talk for 8 hours about free will since there doesn't even exists such a concept as free will. It's very simple: Free will doesn't exists, it's just an illusion.

    1. Re:Free will? by AmiMoJo · · Score: 2

      It depends what you mean by free will. Perhaps physics can show that there is no free will, but that's different to the more philosophical question of if individuals can make free choices and be held accountable for them.

      --
      const int one = 65536; (Silvermoon, Texture.cs)
      SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
    2. Re:Free will? by gweihir · · Score: 2

      All the "proofs" that there is no free will are defective. I have looked, but it was a while ago. In actual fact, nobody knows for sure, but it looks very much like free will exists. Of course, that idea collides with the world-view of a specific type of quasi-religious fanatic, namely the physicalists. Because there is most definitely no mechanism for free will in Physics, so, by their screwed up beliefs, it cannot exist. That Science actually claims no such thing does not hinder them from claiming to have the "scientific truth" in the matter. Just like the fanatic religiots typically claim to have absolute truth with no proof whatsoever.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  3. It is a plot device by houghi · · Score: 3, Insightful

    The first question is about the show, so that is fine. The second is about the future and although it is nice to hear his opinion, it is not more or less relevant than yours or mine.

    I see this also when people quote the three laws of robotics as if they are real. They are not. They solely exist because they can be used to drive the plot. Without them the books would be boring. In fact, the books are basically showing how to get around these laws.

    So please take it as it is, a persons opinion that is just as valid as that from any troll.Because what he is saying is "I imagine that it can be possible." That is good and that is also his job and a way to make a great show possible. That does not mean it is realistic.

    --
    Don't fight for your country, if your country does not fight for you.
    1. Re:It is a plot device by gweihir · · Score: 2

      Have a look at current neuro-"science" research. Much of it is really bad. There are even some neuroscientists that poke fun at the abysmal state of their own fields, for example in "Neural Correlates of Interspecies Perspective Taking in the Post-Mortem Atlantic Salmon: An Argument For Proper Multiple Comparisons Correction", where some of the few good neuroscientists do an FMRI of a dead (!) salmon and find things like that it reacts to voice stimuli.

      Anything coming from this field should be regarded with extreme skepticism.

      --
      Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
  4. Androids will always be merely clever machines. by Sqreater · · Score: 2, Insightful

    "Q: Can we make androids behave like humans, but without the selfishness and violence that appears in Westworld and other works of science fiction? A: I certainly think so. I would hate to be wrong about this, but so much of human behavior has to do with evolutionary constraints. Things like competition for survival and for mating and for eating. This shapes every bit of our psychology. And so androids, not possessing that history, would certainly show up with a very different psychology. It would be more of an acting job -- they wouldn't necessarily have the same kind of emotions as us, if they had them period. And this is tied into the question of whether they would even have any consciousness -- any internal experience -- at all."

    How naive people are. No, we can't. The Human Motivation Array is 4 billion years in the making. And who says selfishness and violence are bad? Not the evolutionary process certainly. They satisfy parts of the HMA and dissatisfy other parts at the same time. They are obviously necessary -- or they would not be there. They would have evolved out long ago. The complex, evolved HMA delineates a behavior-space that we share, - the nominal HMA - but differently accented subtly individual to individual (You can see this on the nightly news, especially the badly maimed HMAs.) You can see this by looking at us. We recognize that we are all human, but we recognize that we all look different. Our entire physicality is our motivation array as humans and as individuals. When you look in the mirror something 4 billion years in the making is looking back. And "Sault's law" (to order my thinking) states that a thing cannot make an artifact as complex as itself. It is an asymptotic goal requiring more and more effort and resources but never reaching the goal - like the speed of light. Why? Because you must know more about reality than the thing you are creating. We cannot know ourselves completely from the inside. Humans will always be able to tell when they are interacting with an android when they grow up around and interact with humans. We communicate each to the other the internal state of satisfaction of our complex motivation array through emotions. Emotions are the state indicators that evolution made for us to interact in groups. Groups are not possible without them. We perceive the internal states of others and react to those states by modifying our own behaviors - and we are motivated to do that if our motivation array is "normal." The HMA will never be replicated in a machine for this reason, we can't see it in detail. It keeps getting in the way of our thoughts and perceptions of reality. Like putting a "colony" on Mars. We cannot bootstrap ourselves. Remember that scientists have said that 100 Billion humans and things that can be called humans have existed. There are seven billion of us today. With the snap of the fingers we will all be gone and replaced by billions more. And more, and more, and more....We are cells in the body of the evolving human species. We are a construction of nature over billions of years. We will not be able to replicate that.

    And I've been recently thinking that our very fuzzy perception of the existence of the HMA is what we call "God."

    --
    E Proelio Veritas.
    1. Re:Androids will always be merely clever machines. by MobyDisk · · Score: 2

      I agree. I expect that we will find there is a spectrum for intelligence. On one end of the spectrum there are brains that are deterministic, efficient, logical, unerring, and unselfish. On the other end of the spectrum there are brains that are adaptive, creative, insightful, error-prone, and emotional.

      For evidence, look at what happens when we try to impart some of that fuzzy intelligence onto computers. They start to make the same kind of mistakes that the squishy brains do. They mistake a rifle for a helicopter or get totally confused by random static.

      The soft squishy animal brains rely on fuzzy logic, statistics, and probably quantum mechanics. They are terrible at multiplying large numbers together, or following a completely logical problem to it's end. They tend to skip steps, or make intuitive leaps to avoid executing each line of reasoning. They sometimes think someone has a gun when they don't. We evolved a brain like this because the world we operate in isn't binary. Does that person have a gun? Well, it's dark, and far away, and they are against a black background, and there are things that look like guns but aren't guns... it's a soft squishy world we live in, and we have to make decisions with partial information. intelligent robots will have to live in this same world.

      Quantum computers also don't find the same solution every time. They use statistical randomness to come to a "most likely" solution. We have to run the algorithm over and over until it converges on what a right answer. The trade-off is that they are very fast, and they can make leaps that simply cannot be done with a pure Turing machine.

      So our current evidence shows that it may not be possible to create a machine that combines the benefits of both types of brains into one.

    2. Re:Androids will always be merely clever machines. by be951 · · Score: 2

      You seem to be making a few assertions here that are simply your beliefs, but using them as facts to support your conclusion.

      • For instance, the notion that we can't replicate something that has evolved over millions or billions of years. (BTW, humans/pre-human ancestors only branched off from other hominid about 7.5 million years ago. The earliest estimates for life existing on Earth are about 3.8 billion years ago, so no, human consciousness was not evolving before single-celled organisms.) However, we have replicated bipedal locomotion in robots, despite that taking considerable time to evolve in our ancestors. So I'm not sure why you think mental processes cannot be replicated.
      • You also claim that humans will never understand human consciousness, but only cite a philosophical bon mot or two "a thing cannot make an artifact as complex as itself" and "you must know more about reality than the thing you are creating" that have a satisfying sound to them, but no evidence that they are actual "laws of the universe".
      • Here's another good one:"We perceive the internal states of others and react to those states by modifying our own behaviors.... The HMA will never be replicated in a machine for this reason." Except that robots and chatbots that observe and respond to human emotion already exist. And evidence suggests that they will be better at it than humans before long.

      It seems to boil down to either "I can't conceive of how it is possible, so it must be impossible" or just "it's a really hard problem", neither of which is a compelling argument to me. There's another piece to it, also, that you may not have considered. You seem to be assuming that people, humans, need to fully understand consciousness and will then need to build it from scratch. However, you're overlooking the possibility that an advanced set of hardware and algorithms that forms a "thinking machine" of some type will develop consciousness on its own. Consider that evolution of organic entities takes a long time because many generations may be needed to fully develop the adaptive traits. Software is much more malleable. It can change in response to stimuli in real time and undergo hundreds of iterations of changes in less time than it takes a person to recharge as is required daily (sleep).

      Machines might never achieve consciousness or emotion that is similar to humans, but it's way to early to declare it impossible.

    3. Re:Androids will always be merely clever machines. by lurcher · · Score: 2

      IMHO, the central flaw to your reasoning is the assumption that we need to understand something 100% before we can create it. We can make a firework (and did) before we understood the chemistry that is involved in gunpowder. Evolution (the process that caused us to exist) is not conscious, it just rolls the dice and then applies a measure (survival) as to the value of the outcome. So the simple can create the (more) complex.

  5. Tags: HBO, Entertainment by NicknameUnavailable · · Score: 2, Insightful

    Really? AI is a valid subject for certain, but this isn't a serious discussion on AI in any regard, it's just viral marketing for a TV show.