Kurzweil predicts we will create software that emulates the human mind by reverse engineering the human brain

By August 18, 2010Ray Kurzweil

This is the third post in a series summarising the key arguments of Ray Kurzweil’s The Singularity is near: When humans transcend biology.  The previous posts were: Ray Kurzweil, The Singularity and the accelerating pace of progress and Kurzweil predicts personal computers with the power of the human brain by 2025.


image Computer hardware by itself is, of course, little use to anyone, so after arguing that we will have personal computers as powerful as the human brain by 2025 Kurzweil goes on to discuss the software that will run on this amazing platform.  His contention, in a nutshell, is that we will reverse engineer the workings of the human brain and emulate the salient parts in software and that the resulting machines will be interact with us in a way indistinguishable from humans.  Further, for all practical purposes, a machine whose interactions are indistinguishable from a human will need to be treated as a human and therefore regarded as conscious, and I guess alive.

It is worth dwelling on the nature of machine capable of ‘interaction that is indistinguishable from a human’, or in other words one that would pass the ‘Turing Test’, of which you might have heard. Such a machine would have to be able to get jokes, infer context, display and comprehend emotion, and exhibit every other human trait you can think of – otherwise it would be distinguishable.  As Kurzweil points out several times in The Singularity, such a machine would be treated as being alive by most people.  Just think how many of us impart human characteristics and personalities to complex but inanimate objects like our cars and computers.

Kurzweil’s contention is that this software will be made possible by continued exponential improvements in scanning technologies (lower power, greater resolution) which will enable us to understand exactly how the brain works.  He is emboldened in this prediction by his conviction that in the next decade or so nano-scale scanners in our blood stream will be able to scan our brain at incredibly high resolution from the inside.  Once we understand how the biological brain works it will be relatively simple to reproduce it in software – a task made easier by the fact that many areas of the brain won’t need to modelled – for example those dedicated to life support and areas where it is sufficient to model at the system or module level.

If you are wondering how far this technology has got today – the answer is the level of a cat brain – which IBM successfully modelled in November last year, and we are also starting to get models for regions of the human brain – e.g. Watt’s model of the auditory regions..  Additionally, many of the underlying artificial intelligence techniques have been successfully operating in narrow domains for some time – e.g. neural nets used in facial recognition, Markhov systems used in speech recognition, genetic algorithms in jet engine design.

This is a good moment to make the point that such a software will not be deterministic in the way that most computer programs are today, but will copy the emergent and chaotic systems of the brain.  The outputs of these human intelligent machines will have random and unpredictable elements, just as human brains do, and they will very much be products of their environment.  The growth and development of an emergent system is hugely dependent on the early environment in which it finds itself (or in other words the early input variables) – look at the way different ways ants nest develop for example.  Different environments will therefore produce very different human intelligent machines.  We might come to think of those differences as different personalities.

These human intelligent machines would start as blank canvases, and to do anything useful they would need nurture and education,  in much the same way as a human child.  One of the reasons that Kurzweil is confident that software will reach this milestone is that each of our human brains is created from a relatively small amount of information – half of the human genome is dedicated to the brain, and that comes to about 25 million bytes, when compressed.  25 million bytes could be represented in around one million lines of code.  Our adult human brains are infinitely more complicated than this and would therefore be much harder to reverse engineer.

Kurzweil’s contention that a reverse engineered machine-baby-brain would grow up to have emotions and all the other aspects that characterise the adult human mind is one that many have taken issue with (including ‘twainventures’ who has left numerous helpful comments on this series of Kurzweil posts).  It is also one that raises philiosophical and religious questions about what it means to be alive or to be conscious, or what it is that provides the spark of life, if anything.

My feeling is that if something (inter)acts like a human then to all intense and purposes it is a human and that questions about the the spark or life etc. are redundant, but that is pure conjecture.  Until we build such a system and see how it interacts I think it will be impossible to know more.

I’m going to close this piece with the observation that once a computer achieves a human level of intelligence it will necessarily sour past it.  Simply put these minds will apply their intelligence to improving themselves and will be able to re-write their own software, very quickly access and upload any information they need to make themselves smarter, and even create new more powerful hardware platforms to run themselves on.

By now you are probably picturing Terminator like nightmare scenarios of machines gone wild – a subject I will return to in the final post of this series.

Enhanced by Zemanta
  • Pingback: Kurzweil predicts we will create software that emulates the human … | SFVN - All for share()

  • Anon

    Often I think that if Turing was alive today he’d have revised his criteria for machine intelligence in the light of these developments:

    • Neuroscience as a scientific discipline became established in the 1960s after his passing in 1954.

    • Following Alfred Binet’s tests for children’s intelligence in 1904 there have been a whole series of psychometric and IQ tests (including Myers-Briggs 1962, Bebin 1981, Baron-Cohen 1985 and more) that redefine what we consider to be “intelligence”.

    The limitations of IQ tests have also been documented, including: cultural bias, left hemisphere weighting and sensory exclusion. Sensory exclusion meaning that most IQ tests are based on our ability to read and visualize the problem rather than to hear, touch, taste or smell it.

    So, for example, is there intelligence in Nigella Lawson being able to smell variations in chocolate depths and notes or intelligence in Roger Federer being able to hear and “touch” the weight of a ball on his racket and determine how much topspin is needed?

    Notably, neither of these skills is tested in IQ tests and yet both people are intelligent.

    • The invention of MRI scans in 1977 transformed how much we can see into the brain. Plus the advancement of nuclear medicine in the 1970s, which enabled most organs of the body to be visualized.

    (NB: visualizing it is not directly equivalent to experiencing and walking in every step of its myriad of interactions.)

    All of these advances have changed and are changing how much we know about the human brain and about the nature of intelligence — knowhow not available to Turing in his time.

    If he was alive I don’t doubt he’d propose that instead of Babbage’s difference engine (which itself forms the basis of the Enigma machine), we should be examining ways to develop coherent differentiation systems in the first instance and use these as a basis of then building out context and eventually artificial consciousness — again the distinction should be kept between organic human consciousness and silicon machine consciousness.

    Just because we attribute (rather than impart) “human characteristics and personalities to complex but inanimate objects like our cars and computers” does not mean they are alive. They’re not oxygen, water and sunlight dependent for their existence and growth so even by this most basic of definitions of Life………….they cannot be alive.

    These inanimate objects are a reflection of our intelligence rather than intelligent in their own right; the machines depend on us humans to reproduce and to adapt their shape, materials, personalities — “A customer can paint a car in any color he wants it as long as it’s black” as Henry Ford said and we’re all aware that human marketing is responsible for giving objects like the iPad identities and characteristics rather than the object being able to self-generate these.

    Consider also the issue of injuries and viruses. Organic matter which is alive strives to stay alive and regenerates tissues and chemicals that can help that organic matter to ward off or deal with that injury or virus. An inanimate object that gets broken or catches a virus stays in that state and is dependent on human intervention, repair and troubleshooting. That’s another obvious distinction between what constitutes something being alive and something being an attribution by something that’s alive.

    Now it would take too long to dissect Turing’s Test in detail and there are people more intelligent who can share their attempts to pass the Turing test with their machines:

    * http://blogs.discovermagazine.com/sciencenotfiction/2010/06/28/watson-fails-the-turing-test-but-just-might-pass-the-jeopardy-test/



    What is important, though, is to be reminded of the specific premise and parameters of Turing’s paper on ‘Can Machines Think?’

    At its core it’s about a machine MIMICKING a human purely according to text-based inputs.

    It isn’t about a machine being able to interact as a human at all.

    For example, the Turing test doesn’t require a machine to do face recognition to determine whether we’re smiling or showing any other emotion and the extent of authenticity of that emotion from our body language and pupil dilation. It isn’t about the machine being able to smell our hormones to determine other emotions (fear, desire, anger etc.) either. Nor is it about a machine that can hear the emotion in the timbre of our voices or that can reach out and touch our smile / tears out of sympathy or empathy.

    Importantly, it isn’t a test for Consciousness as such. It’s a test to see if a machine can copy our pattern of text-based conversations (understanding and extracting the context, humor, emotion etc. from the text).

    There’s simply so much more complexity needed for a true test for Consciousness and we haven’t even definitively answered what that is for the human brain yet…………….

    So…………… to the books, the drawing boards, the brain imaging scanners, the Web and the code kitchens for all of us!

  • Anon

    Wrt “Once we understand how the biological brain works it will be relatively simple to reproduce it in software….”, it may be helpful to read this book because it may not be sufficient to “model it at the system and module level”:

    * http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=11003

    * http://mitpress.mit.edu/books/chapters/0262162393pref1.pdf

  • I am now o;d enough te remember AI people saying we would have computers that matched the Human brain25 years ago 🙂

    Anyway, its all going to be run by Google – http://bit.ly/92Y6cK

  • One day the predictions will be true 🙂

    And it would be a fool who bets against everything that Kurzweil says. Plus, you know me, I’m an optimist.

  • Pingback: Finance Geek » Kurzweil on coming revolutions in genetics, nanotech and robotics (strong AI)()

  • You haven’t ever said this, but it seems to me you might believe that there is something very complicated and possibly spiritual about consciousness – or the process of becoming conscious. Am I right here?

    My view is that Occam’s razor prohibits such a position.

  • Anon

    Spiritual? No. Emotional, socio-relational and scientific? Yes.

    Let me provide some context: my father passed away following a coma.

    The neurosurgeons insisted he was “unresponsive” to their established tests for consciousness (ECG, MRI, physical mechanistic stimulus), so by their definitions he was unconscious. The hospital later wrote a standard letter to the coroner to state this and, as far as they were concerned, that was that.

    After I read the letter I sent the coroner video evidence that showed my father was conscious and to encourage the hospital to explain what had happened in coroner’s court.

    The result was that the hospital and the lead neurosurgeon apologized to my family in coroner’s court.
    My father had shown signs of consciousness — just not as defined by the existing medical tests or instruments available to map his brain activity.

    Sadly, he would have passed away regardless because of complications relating to his organs. Nonetheless, what happened demonstrated that we (including the medical profession) do not yet know enough about the brain and the what-where-why-how-whens of Consciousness.

    Maybe it’s simply not measurable via ECGs, MRIs or pressing down on someone’s collarbone to see if they respond.

    Maybe we need a whole other set of instruments.

    It’s like this: before the invention of weighing scales we had no way of measuring weight (mass, m). If Newton didn’t have scales to measure the mass of his apple…… even if he had been able to calculate gravity (g) to be 9.8 ms-2 we would now have no constructs for Force (F) since:

    F = mg

    So no weighing scales means……….No Newtonian force => no thrust to overcome gravitational pull => no steam engines => no automobiles => no aerodynamics => no Wright Bros => no Einstein’s Quantum theories => no nuclear bombs => no stealth planes => no Concorde => no NASA => no Stephen Hawking advising us to colonize other solar systems before the aliens arrive and eradicate our species — that is if we haven’t already been exterminated by the AI Terminators and Skynets that we could potentially build or wiped out like the mammoths in another Ice Age because of our consumption impact on the climate or annihilated each other with nuclear weapons because as Hobbes observed “the life of man is solitary, poor, nasty, brutish, and short” or some other catastrophe like an asteroid collision or the sun falling out of the sky etc.

    I digress but…..

    Someone had to invent the weighing scales.

    In the same way………someone has to invent the tool(s) that will enable us to capture Consciousness as an entity (whether it’s a force in Newtons or energy in Joules or an electrical conductivity in Watts or some other scientific metric remains to be proven).

    How does what happened wrt my father’s consciousness relate to Web technology and the Singularity?

    Well, since the neurosurgeons couldn’t explain consciousness in a way that made sense and since I’d observed consciousness online (brand, issues, political movements and more) I tried to synch the two worlds — medical and Web — to construct my own consciousness molecule, which has the following constituents:

    • cognizance

    • coherence

    • communication

    • creativity

    • command-control

    • consideration

    • collaboration

    • culture

    So unlike Kurzweil whose postulations about the genome and Moore’s Law have already been called into question by scientists and computer programmers alike:

    * http://news.ycombinator.com/item?id=1611007

    * http://www.sentientdevelopments.com/2010/08/myers-kurzweil-is-pseudo-scientific.html

    * http://www.spiegel.de/international/world/0,1518,709174,00.html

    my approach to consciousness — and, by extension, potential code frameworks for artificial intelligence — are not based on the metaphysical and mechanistic logic functions of the brain alone.

    In my model of artificial consciousness, the logic functions correspond with command-control. This would be where the binary and probabilistic algorithms apply. Communication and collaboration are emergent in IM, Twitter, Facebook whilst cognizance can be found in face, voice and smell recognition technologies. Cyber culture is already a given and growing its own identities via tags, like buttons and community dynamics (ranging from influencers to trolls to lurkers).

    What’s clear is that code still needs to be innovated for coherency, creativity and consideration (the emotions, morals and values bytes).

    As I noted previously neither I nor anyone else has the definitive answers; we’re simply going on expeditions for Consciousness according to whatever informs our experiences.

    Kurzweil believes the code answers to the artificial brain lie in genome. Interestingly, Craig Venter the man who successfully sequenced the human genome and is involved in all types of experiments involving it ranging from biofuels to software gave this interview to Der Spiegel in July 2010:

    * http://www.spiegel.de/international/world/0,1518,709174,00.html

    Meanwhile, my father’s passing was like an apple falling from the tree onto my head. It’s a different apple from Newton’s, a different one obviously from Kurzweil’s and any other person’s apple.

    Each apple is a challenge.

  • Anon

    Ockham’s razor is the MBA mantra of KISS from another age (Keep It Simple, Stupid).

    That’s fine if we’re talking about UX, the iPod or a business model (especially Google). However even there, Ockham falls apart. Beneath the seeming simplicity of that single search box is a complicated series of complex algorithms with lots of lines of code, data farms, libraries etc.

    Moreover, as we’ve seen……Google’s seeming simplicity doesn’t mean it’s got the best answers — just the most frequently PageRanked. That’s why a range of semantic search companies decided to provide alternatives they considered to be better than Google’s.

    As for humans…….we’re complex by nature. It’s just that the myriad of nano-scopic interactivity within us isn’t completely visible nor mapped (yet).

  • Occam’s razor (interesting you can spell it either way) has it that all things being equal the hypothesis that introduces the fewest entities and introduces the fewest new questions/assumptions is the better one. On this basis I can recognise the complexity in everything you describe, but reject the notion that there is something indefinably spiritual about consciousness.

  • I just got to this comment – sorry to hear about your father.

    I guess I think it is entirely possible that consciousness is identifiable and measurable, but we don’t have the instruments or understanding yet, but I also think it is possible that if we create a computer with the same logical functionality as the human brain and train it as you would a child then it will become conscious. Given Occam’s razor I think the latter is the better explanation. The way you break down consciousness and identify the remaining areas where new code is required leave me thinking our positions are actually quite similar.

    I agree that it is impossible to know at this stage though, and hence my position is pure speculation. I also recognise the huge complexity in the human mind and the current paucity of our understanding. I believe quite strongly that we will chip away at the gap and over time we will close it.

  • Anon

    Thanks and one of the outcomes of my father’s situation is a determination to highlight that more needs to be done (medically and in the wider world) to understand Consciousness, possibly locate it and then harness it constructively.

    Also, since consciousness is not yet fully understood by the medical profession it means that coma patients may be suffering unnecessarily. Imagine a young person who’s involved in a car accident and the doctors diagnose they’re unconscious due to lack of ECG / MRI activity when in fact they ARE conscious — just not measurable with the ECG / MRI; this consciousness not being based on their electrical activity, blood flows and functional physiology alone.
    This would change how moral decisions are made about whether and when to turn off the life support systems.

    Next, I appreciate your intent wrt “chip away at the gap and over time we will close it.” You mean that human endeavor and tenacity will triumph. Yet, paradoxically, when we literally “chip away at the gap” they actually become wider because the empty space in between becomes bigger.

    That’s an example where Semantic technologies aren’t (yet) able to properly capture meaning, intent or consequence btw because Natural Language is unable to extract whether our sentences are figurative, literal, paradoxical, ironic or referential.

    Nonetheless, I do share the belief that through collaboration — and most likely on an open and multi-disciplinary basis — we can conjugate scientific approaches (including neuroscience, Quantum Theory and DNA), computer code, linguistics, socio-psychology and economic models to derive solutions that might get us greater Consciousness in our technology.

    There’s an obvious variation between my approach and what Kurzweil propounds about the Singularity. I start from the basis that humans are more intelligent and complex than machines; also that the heritages of our DNA as well as our moral conditioning affects our socio-economic-philosophical psychology and human contracts.

    Meanwhile, Kurzweil seems to want us to revert our brains to mere matter (information) that can be transmitted digitally in the same way that information was via the Reading Machine he invented.

  • Pingback: We Will Create()

  • Anonymous

    Top read. If they did manage it why would anyone need to employ humans anymore; if say these robots could go without sleep or food?. I think that they could never build something that could creatively think.

    Document Scanning

  • Pingback: Kurzweil part 5 – the impact « The Equity Kicker()

  • Pingback: Scientists create artificial brain with 2.3m simulated neurons « The Equity KickerThe Equity Kicker()

  • Pingback: The difference between brains and computers « The Equity KickerThe Equity Kicker()