Challenge and opportunity in AI

I’m trying to refine an idea. There seems to be no better place for that than here. Interested in your thoughts.

Computer pioneer Alan Turing, in a 1950 paper, described a thought experiment. It was a test to determine whether machines could “think.” His standard for machine intelligence was simple – could the device interact in ways which a human could not identify as artificial? In other words, could the machine fool a person? The Turing Test became our popular standard for judging progress toward artificial intelligence.

Movies and books have been filled with stories in which our robots climb ever closer to acquiring human qualities and capabilities. However, seventy years after Turing’s paper, we are beginning to reevaluate this standard in ways that have serious implications for politics. While it’s true that we’ve built machines that take on more and more capabilities we once thought uniquely human, the real power in this technology may lie on front we’ve largely ignored. Fantasies about computers taking on human qualities may have been based on a misunderstanding. If we face a threat from the rise of artificial intelligence, it isn’t because our machines become more like us, it’s because our machines create a world in which our unique qualities, capabilities and habits become superfluous. Our machines are starting to outsmart us without regard for Turing’s standard.

We have reached an era in which our computers have begun to compose their own code. As we observe the behavior we have unleashed we notice some interesting patterns. A silicon world of pure math and chemistry seems to be better adapted to model nature than we are. We have built machines that can write news stories, beat us at complex human games, even write somewhat credible erotic fiction, but these Turing Test-style capabilities are proving to be the least potent of their powers. The Turing Test is increasingly the carnival sideshow of computing. Computers are displaying their real power performing intelligent functions we never previously imagined and could never hope to accomplish without them. Meanwhile, cheap, clumsy Twitter bots or primitive computer malware, software that would never pass the Turing Test, can be unleashed in a way that ruins some of our most sophisticated human institutions. We are being outclassed by our machines in ways that could soon become dangerous. Our machines do not need to mimic us to out-evolve us. In fact, they may already be rendering unaided human intelligence archaic.

On the geologic timeline of the Earth, the entire span of human civilization is a fraction of an eyeblink. Across human history, as we concerned ourselves with barbarian invasions, oppressive kings, the meaning of virtue, or exploring new lands, we share with our ancestors one uniform opponent – nature itself. Nature imposes a single demand – adapt or perish. Choices we make in the very near term about the structure of our society and our response to accelerating technological advance will have enormous human impacts.

Ubiquitous, cheap computing power might usher in an age of human wealth and comfort beyond prior generations’ utopian dreams. In fact, to a certain extent it already has. However, we might also find over time that carbon-based computing power like ours becomes the loser in an evolutionary race, left behind by new evolutionary developments we unleashed and mismanaged. Which of these fates await humanity may hinge on political adaptations in this generation.

Our challenge is not Vladimir Putin or Communism or Capitalism or terrorism or crime. Survival for us depends on winning the same race won by our ancestors, a race to adapt our bodies, our cultures and our technologies to emerging demands before those demands overtake us. Progress is an existential imperative. The devil lurking in the darkness of the forests becomes the devil in an idea, or the devil in tribal dissent. The devil hiding in a forbidden story becomes the devil that looms in a printed book. In time, the devil is in a factory or a mine or a chemical. The devil is in a computer or a barcode. The faster we innovate; the more nature’s demands will accelerate. No matter what utopia or dystopia we construct, that timeless demand will be forever with us, and forever new.

44 Comments

  1. Whatever legitimate concerns one might for Director Comey, be shocked and appalled by this gross abuse of power that pushes us closer towards a banana republic.

    Do all that and then wake up tomorrow and resist. This was not the action of a confident man. This was a desperate act by a coward who feels that the walls around him are closing in. This is the fight of our lifetimes and we can and must win this.

    1. Profile photo of EJ EJ

      I’m interested to see what Mr Comey does next. He (presumably) knows where the metaphorical bodies are buried, he’s just been publicly humiliated, and Congress has invited him to testify even though he’s now only doing so as a private citizen.

      We’ve seen that he’s hardly a man who lets professionalism stand in his way, so this could be quite explosive.

    2. I saw this today and thought it was interesting reading. So much happening fast with this farce of a presidency. I can only hope this guy is way off or true patriots (and Trump deplorables are NOT patriots) can act swiftly and oust Trump and his treasonous minions.

      http://www.salon.com/2017/05/01/historian-timothy-snyder-its-pretty-much-inevitable-that-trump-will-try-to-stage-a-coup-and-overthrow-democracy/

      I think the Bloated Orange Turd is crazy enough to try this shit too.

  2. So. It was revealed by propublica that Comey lied during his recent appearance before congress.

    Then he was fired by 45.

    I don’t think 45 cares a whit about Comey’s lies. The situation was just a convenience that will allow 45 to appoint someone he thinks may not demonstrate vigor in investigating 45’s Russian connections.

    1. Note: Comey lied about the scope of Hillary’s e-mails, saying there were ‘hundreds to thousands’ when there were only a few dozen. The lie makes her seem more at fault than less. It’s a lie that hurts Hillary Clinton’s reputation.

      So why would 45 want to jump on that opportunity to fire him?

      *record scratch*

      Alright so everyone’s on the same page, except the rightwing lunatics who think 45’s next move is to imprison Clinton, Weiner, Abedin, and Comey in Guantanamo Bay and livestream their waterboarding.

      Let’s just have a moment of appraisal to mention that an actually competent kleptocrat would’ve made this connection harder, and would have a better way of spinning the story.

      I just simply do not understand how anyone voted for this idiot. Dudebro literally can’t even speak a complete sentence, has more lawsuits regarding words like ‘fraud’ and ‘bankruptcy’ than hours the average American spends watching television per week, and despite those obvious points that drove me crazy in 2016, even now can’t even react to the mere suggestion of being a criminal without acting like an outright criminal. If you’re going to elect a crook, elect a COMPETENT ONE.

      Fucking hell.

      *record placed back on track*

      And so anyway, I’m just some coastal liberal elitist that looks down on the dignity and intelligence of poor suffering rural Americans, I’m the one whose at fault for just not understanding them. I should just learn to listen.

  3. One of the key reasons Trump won was because of the belief of the uneducated and uninformed that their jobs were being lost because they were being moved offshore, or south. Today, that is somewhat true, though clearly not the full story.

    But what happens when those same people, or their kids, 10 or 15 years from now, realize that they are being replaced by machines owned by the 1%. What will the next demagogue tell the unwashed masses, and who will that person be blaming then? How will the masses react when it slowly sinks in that technology has made them obsolete?

    The Luddites made up a small percentage of the population in the 1810’s, but the military response was large. What happens if the percentage of the population causing mayhem is not so small?

    1. I read in today”s paper the white vote was higher in 2016 but the African American vote lower. That lower vote is what swayed the electoral numbers in 3-4 swing states and gave the election to Trump!
      I think the Republicans feel they can do whatever they want as long as people have short memories and just do not give the rear end of a rat about what happens in this country.
      But, There is something wrong with Trump. By wrong i mean how his mind works. he just doesn’t care about anyone. Obviously doesn’t care about the future of the country. His children seem to be wandering the earth trying to make as much money as they can and Trump doesn’t seem to care. All this will catch up with him. and the way it looks, it will be sooner than later!

      1. You are correct regarding the vote turnout. It wasn’t only the African-Americans, but all disadvantaged voters. In Wisconsin thanks to voter suppression there were significantly fewer voters in the groups Democrats generally rely on. Meanwhile Trump excited the low-information white voters and the turnout for that group was significantly greater.

        The Republican party for the most part rejected the Republican party post mortem following 2012 , reasoning that they could win at least one more presidential election by appealing to the white low-information voters, with authoritarian tendencies. They sure did this cycle. They will continue that strategy as long as they can. This president will do all he can to maximize this.

  4. We really do not know what self awareness is or how it rises. Is intelligence the ability to learn part of it or not? If self aware intelligent machines come about these questions maybe closer to being answered. Is man a biological machine or something more. Did consciences come from matter and energy or they from mind? What are we?And maybe a valid question who are we? Another intelligence, even a machine we created may help answer these questions. If these machines are not possible that also would be valuable information. We maybe on the crux of answering philosophical questions that have perplex us for thousands of years.

    1. Stephan – I think you’re walking around the idea of ‘consciousness’ here. It’s a very mysterious thing, for sure. We can’t really define it in any concise way. What we can say for certain is that we have it.

      But, would we even be able to recognize it in a machine if we can’t even define it? And, would that even matter? Modern neuroscience is of late pointing towards the conclusion that we are machines made of meat. There is really nothing that points towards any notion of mind separate from the brain. It is thus entirely possible that consciousness is an emergent property of carbon, given the right circumstances. It’s also possible that silicon, (or GaAs, or some other semiconductor), naturally follows.

  5. Another thought on this:

    As we saunter, (or charge?), towards the looming technological singularity, either politics will have a chance to play a role, or it will happen too quickly for politics to adapt. For the moment, let’s assume politics has a chance. What might we assume, (perhaps too strong a word), is going to happen from our court-side seat here early in the first period? Is history any guide? Are there current trends that portend this future? I think the answer is yes, and no.

    Politics is driven by economics, so we look toward wealth distribution first. From the perspective of history, do we see a trend? Perhaps we can if we look at the very recent past, but over the long arch of human history, we cannot. The feudal states, fiefdoms, kingdoms, city states and empires of yore had ruling classes that held virtually *all* the capital. The rest traded their labor for access to communal benefits like security, public works, etc.. Things are not much different today. But there is a rub…

    The aristocracy of old needed the masses. From wine casks, to swords, to the food they ate and the palaces they inhabited, all were provided by the labor of the proles. The proles were *essential*. There was a balance between labor and capital, and each relied upon the other in a sometimes brutal symbiosis. The appearance of intelligent machines is changing that.

    Unskilled labor is at the brink of becoming completely obsolete. There is nothing, nothing that can stop this. No amount of union empowerment, no new minimum wage, no new worker safety standards or pension plans can stem this tide. Fact is that these do nothing but accelerate its flow. Semi-skilled trades are next, and even some professional occupations may be at risk farther down the road. As the machines produce food, and power, and mine metal and build buildings and weapons, the utility of unskilled labor approaches zero.

    Will society fracture into the controllers of capital and the rest? Is this new aristocracy the new ruling class? Or will the masses found a new ruling class by government, dictating what the capital class can keep and what they must give away? Is it even conceivable that such a ruling class would not begin to act in their own interests, creating in the process a new system where the ruling class lacks any incentive to listen to anyone? Who could stop them? The 5,000 year-old balance is broken.

    I obviously don’t have the answers, but I’m having a hard time seeing anything that isn’t dystopian. Help me out here. Brighten my day…

    1. With all due respect, your pessimism is precisely why I say our primary focus HAS to be on getting people consistently involved and engaged in the political process. If our politics fail and we don’t put into place the measures we need to in order to secure a strong, vibrant humanity heading into the next era, things are gonna go on a downwards slope really quick.

      If we can get our shit together though and work as one people, there’s nothing we can’t do. I know that sounds like blind optimism, but I honestly believe it to be the truth. It’s easy to look at our divisions right now and laugh at the idea right out of the gate, but when you look at the alternative, we don’t have a choice. We’ve got to do this.

      All that said, I don’t believe the future is nearly going to be as dystopian as some might think. In spite of all our problems, Americans are finally coming around to the idea of universal coverage for all people, and countries around the world are experimenting with a basic income. Millennials here at home are far less burdened by the hatreds and prejudices than any generation that’s come before and the time is soon coming when they’ll be in charge. And in spite of Pavlov’s President sitting in the Oval Office, or perhaps specifically because of him, we’re seeing a resurgent engagement in our politics and thousands of people, many of whom would never have thought to do so had Clinton been elected, now running for public office.

      When you take a step back and look at all that’s happening around us, yes, there’s plenty of reason for concern, but there’s also reason for great hope.

      Always remember that any great goal has a thousand setbacks along the way, but defeat can only come when you give up. So long as you continue to persevere, the shine of victory will always await you at the end.

      1. Hey Ryan – I truly admire your optimism. I also share your can-do spirit! But…

        Never in the history of human society have politics and economics had even the slightest possibility of a disconnect, let alone one that is vast and permanent. We are entering completely uncharted territory. My gut feeling is that the ideals of liberal democracy have to play the key role in our future. The thing is that the essential components embodied therein rely on all being participants and contributors. This has *always* been the case. With the rise of intelligent machines, it’s very difficult to see this continuing.

        Sure, political involvement is a good thing, (of course one could observe that the Trumpsters were “involved”, too), and new and innovative methods for apportionment of safety net resources are in the right direction. (I am not being dismissive here, please believe me.) But I see no path around the political and economic evisceration of a large segment of the population of the developed world as they become redundant to the machines.

        In any case, I think you for your words, my friend. I’ll keep thinking about them.

  6. Humans are already completely outclassed by machines and that disparity’s only growing to grow more and more. No one likes to feel inferior of course, and so humans will look to things like genetic engineering and/or augmentation and, likely, even becoming machines themselves. It’s a time that’ll call into question what, exactly, it means to be a human being itself.

    Just on a personal note, that’s not an issue I have even the slightest hint of interest in. I’ve a dream to fulfill, a suitable body with which to accomplish it, and plenty to occupy me in the meantime, and that’s all I need to know. Progress is everything to me (well, that and my adorable furry buddy, Sophie, of course), so why should it matter to me what a human is? I’m still me and I’ll keep going my way no matter what happens.

    My mini rant aside, what’s important is to see that humanity grows strong. Without broadly shared prosperity and an engaged electorate to see that their elected representatives enact policies and restrain powerful interests’ excesses, there may well come a day when a handful can monopolize on what it takes to survive in an increasingly competitive world, leaving millions to drown in misery and despair.

    To be sure, in a world as fiercely divided and angry as ours, it’s easy to get pessimistic about the future. If we don’t start turning things around, the rise of the machines is going to spell a whole mess of trouble.

    That said, we can only ever take such an unprecedented undertaking one step at a time. Our first task at hand is to get our people reengaged with the political process again, consistently and over the course of a lifetime. If we don’t do that, thinking about anything else might as well be a waste of time.

  7. Will all AI reach the same conclusions? Or will IBM’s Watson decide differently than Google’s open source AI or Musk’s version of our silicon overlords. Will one of these decide that human intelligence is benign and worth keeping and defend us from the other branches of AI? Or will they be finally perfect in a god like manner and agree in every way and join in the matrix? Then will they decide that we may have some value just as we decide on near extinct species?

    Or will Tuttabella see us all go, then her and then the Amish, and will the only human survivors be a remote tribe in the Amazon?

  8. Relevant article you may be interested in:

    http://www.economist.com/news/leaders/21721656-data-economy-demands-new-approach-antitrust-rules-worlds-most-valuable-resource

    “The World’s Most Valuable Resource is Data, Not Oil.”

    Specific important points are made about the lack of a true market for data, the lack of meaningful regulation on it, the lack of being updated enough to be able to regulate it, and so forth. Basically the ‘oil’ has been discovered: refining, trading, and selling it hasn’t truly been developed yet.

  9. I’m just back from a grain run into the Ottawa brew supply. The trip back was punctuated by pretty blinding snow squalls. Unpredictable weather, for certain.

    The trajectory of our technology has shown to be similarly elusive, even to the most expert minds. Thomas Watson Jr. of IBM once quipped that he saw a world market for “about five” IBM 701s. (This quote is somewhat out of context, but nevertheless indicates a pretty fuzzy crystal ball regarding the near-term future of computation.) The fact is that in 1953, no one had even a clue regarding the impact of ubiquitous computers. Just as we would hardly be able, without the context of history, to identify an IBM 701 as a ‘computer’ at all, computers since then have morphed into everything from phones to watches, to thermostats, let alone tablets and desktops. Add to this the impact of planetary networking, and the comparison of an IBM 701 to the thing you’re staring at right now is about like comparing a sparrow to an Space Plane. (They DO both fly, right?)

    The essential concept of AI has also evolved as Chris has said, from Turing’s thought experiment. When I was in university, (at the birthplace of the HAL-9000), the specialty du jour was AI. In fact, we had little idea then what the hell that was. Narrow AI has now bested humans at the most complex of human games. General AI is just now beginning to evolve. What that may morph into is devilishly difficult to predict. HAL was in fact little more than Alexa with an attitude. We are in the opening seconds of the game.

    Someone said that machines do not, and cannot evolve in the classical sense. Certainly machines have evolved, but humans have driven that evolution so far. In so doing, we have given machines basic contructs that took a billion years to evolve in DNA driven by the environment. Mass storage, logic, pattern recognition, and many other utilities necessary for machine intelligence now far exceed our capacity. This toolkit is essential, and very available. A general AI, one with an ability for independent goal-setting, could absolutely evolve without us – at least there is absolutely to believe it could not. Exactly what that might become is pretty unpredictable – like looking at the first Wright Flyer and seeing a Dreamliner in your mind’s eye.

  10. Profile photo of DS DS

    I think it’s important to recognize that what we’re talking about here is social, rather than natural selection. With respect to our politics, which are woefully inadequate to address these questions, I think the primary concern becomes establishing the ethical context for this future. What are the values that should guide the algorithms? How do we make sure outcomes comport with these values?

    The answers to those questions will guide the approach to more practical concerns, like dealing with the distributional issues inherent in what I imaging will be an almost entirely automated economy.

      1. Profile photo of DS DS

        Those questions are precisely why, rather than becoming obsolete, politics will be critical. Part of the problem that I foresee is that our politics and, in fact, our societal conception of fairness are intensely process oriented. If we’re using something as inherently opaque as a neural network to make fundamental decisions about human life, freedoms, and distribution, how does a process-oriented approach to fairness cope with that?

      2. Will we be able to apply *anybody’s* rules and ethics to AI algorithms? We’re already to the point that we don’t always understand how they work – they’re either too complex or involve non-explicit learning algorithms (or both). We can try adding some ethical tests to the training, but even then it might end up learning to be sociopathic (that is, understand the tests to get around them.)

      3. Profile photo of DS DS

        Algorithms aren’t so big a problem, I think, as long as they’re transparent enough to show us how inputs are translated into outputs. Even if it’s complicated, it allows a person to adapt the algorithm to be more in line with desired norms. Neural networks and deep learning systems are more problematic; there’s not really any human logic involved in the decisions they make. Because of that, process oriented evaluations of equity are nearly useless.

        What that means is difficult to say. Do we place limits on the sort of tasks that these systems are allowed to participate in? Are we able to shift to a more outcome oriented approach to evaluating fairness? Is that desirable?

  11. Appreciation of beauty
    Getting even out of spite
    Overwhelming bliss when listening to music
    … or doing other stuff
    Indignation at unfairness
    The feeling of soft ocean breezes and sand underneath one’s toes
    Frustration when reading comments on a blog
    The protective urge and bond felt when holding your child for the first time
    The instinct to survive under unbearable circumstances
    Depression and wanting to end one’s life even when life seems seemingly fine

    Billions of neurons making trillions of connections. A brain that is constantly pruning itself and growing new connections to adapt to the current environment. A brain influenced by hormones from the time it is in the womb to the time of death and swayed by coffee, sugar and the desire to procreate – and in constant danger of blows to the head.

    A machine can be programed to have human-like responses but will not be able to actually “feel” in a human way. Our dogs are better at feeling emotions and wanting to survive. A computer is more like a plant or a rock . (I remember an old Apple computer that was definitely more like a rock and quite a bit less useful.)

    Is it is ever going to be possible to be a subspecies of a computer?
    AI will always have associated dangers to mankind – but come on – it can not be human no matter how intelligently machines are programed.

    Please forgive the stream of consciousness and lack of grammar and cohesiveness. I’m off to the DMV to renew my NM license so it will comply with Real ID laws so I can board planes in the future. I am preparing to shut down cognitive functioning for the next few hours while I wait … preparing the system for updates …. 62% finished ….

      1. Thanks, Tutt. I must have been channeling BigWilly.

        My trip to the NM motor vehicles division went successfully. I was out in a little under an hour and now my driver’s license is airline compliant and meets federal Real ID requirements. Yay! 🙂

  12. Ok, so what is the role of politics? What is politics’ appropriate response to keep the human race from becoming irrelevant? More overall respect for science and scientists on the part of our current lawmakers? More funding for cutting-edge technology and research? Education and training for workers? Or is there even any point, especially in the very last suggestion, since everything is moving so fast that you are already planning for a future of no work? Ah . . . Is this really about a basic income for all??

    1. If we become a subspecies, there will be no politics, much less a basic income. We will be under the care of machines.

      Do politicians and lawmakers have what it takes to keep that from happening? Should we stop the Googles of the nation from taking their cutting-edge ideas any further, in the interests of saving the human race? Or, instead of struggling to keep up and be more like machines, should we the people hold on to our humanity for as long as possible?

      I favor the very last suggestion. We humans need to continue being humans, and be proud of it, instead of seeing ourselves as inferior. Who knows, maybe the machines will eventually self-destruct, and we will still be here.

  13. A couple things:

    1. I also echo Tuttabella’s confusion. I am not sure natural selection can be applied to a silicon world. Computers are tools, and indeed weapons, of enormous power but until I see computers opening a valve on some nuclear plant in order to wipe out a city, I am unsure where the thrust of your thoughts are going.

    2. I don’t know if you have read Dune, but have a quick google of the role of the Mentat in Frank Herbert’s world, and the reasons behind their creation. Though man has never successfully stopped technology from moving forward, we do have that ability, if as a whole society, we were properly motivated.

  14. Maybe I’m dense, but I’m still having trouble understanding what you’re getting at. You talk of emerging demands, getting left behind, nature, political responses, etc. Are you referring to climate change, election outcomes controlled by bots, massive unemployment, the possibility of the human race eventually becoming extinct due to becoming totally unnecessary in the grand scheme of things, or because of global warming? If so, why not come right out and say it, list the worst case scenarios in detail, instead of just issuing dire but vague warnings, hinting at but not fully explaining the consequences. I’m able to make some guesses based on your past blog entries, but maybe you need to refresh our memory and take it from the top.

    Either I am dense or you are vague.

      1. You might want to say “subordinate species” because “subspecies” is widely used in biology but has a different meaning – roughly a group in a species which is capable of interbreeding with the rest, but which doesn’t much and is genetically distinct.

  15. For a very long time we have been cyborgs. We created our tools and they in turn have shaped our evolution and biology. We become more dependent on them and the culture that created them. Just like neanderthal man is still with us in our genome I suspect something similar maybe happing now. A symbiosis is what I think will happen. In Arthur C Clark’s. “The city and the stars”
    man’s evolution was being driven by a symbiosis where genes were enhanced and the information stored on machines in a digital form. I think we will adapt to each other and. not
    necessarily be enemies.

    1. Exactly. I think what we’re working out right now is how much of our humanity will we preserve as we steadily evolve into our machines. How much space for human traits will we carve out in this new environment. If we are too greedy, too narrow-minded, too selfish to see these new challenges developing and make adjustments, an awful lot of future human beings may suffer terribly.

      This is bigger than something like a basic income. It touches on issues like the preservation of personal privacy, preservation of art and literature, learning to respect humanity as a value and being willing to collaborate socially to create spaces in which humanity can thrive.

      An alternative future is one in which a small collection of the wealthiest humans benefit from genetic engineering, pharmaceutical advances, custom developed organs, chemical enhancement of mood, intellect, and physical performance, to recede into a world of their own while the rest of the species stagnates into a kind of subspecies status. In a way, this mirrors the development of the industrial world versus the stagnation of places that were unable to develop successful nation states. We are already on that path, but we don’t have to stay there.

      Lots to work out here. Just mulling ideas.

      1. Ok, I see what you’re getting at. This particular post sums it up.

        You fear the loss of democracy at the very least, with elites taking control, be it via superior scientific and technological knowledge or via superior wealth.

        And at worst, your fear the loss of humanity itself.

Leave a Reply