The Technium

Expansion of Free Will

[Translations: Japanese]

The evolution of the technium is self-directed. Over time it unfolds a sequence of self-organizing forms. Because these self-organized forms are inevitable, we can prepare for them. But the inevitable aspect of the technium provokes resistance because it appears to be opposed to human free will.

The author Isaac Singer once declared, only half in jest, that “we have to believe in free will. We’ve got no choice.” That kind of mad desperation often crops up in the almost reflex denial of directionality in technological evolution. “We have to reject an inherent direction in the technium because it reduces the sacred role of humans to decide our own fate.”

But what does science say? While the evidence for evolution’s self-direction is only circumstantial today, the claim of directionality is testable one way or the other.  We might be able to prove that biological evolution is self-directed by employing synthetic biology to induce life’s self-assembly numerous times in the laboratory and measuring how often parallel evolution occurs. Or, someday we could scout the galaxy for other planets with life, and when we find living systems, we can tally up the degree to which those alien evolutionary paths parallel our own. These findings would be falsiable evidence. If science did prove that biological evolution was self-directed along a certain universal trajectory, we would not reject that fact just because human choice had no impact on life’s direction. The scope of human choice and evolution’s direction are independent assertions. Yet if science gains evidence that the technium, which is a quickening of evolution, is also self-directed by its own emergent order, humans want to refute this fact because they believe that directionality in the technium denies human choice. But whether it diminishes the scope of human choice should not influence the facts of whether the technium is self-directed or not.

In fact, the worry about diminishing human choice is misplaced. Free will is not hampered by a technological imperative; instead it is expanded by it. The inherent imperative merely shifts the venue of our freedoms. Once inexpensive communication systems circle the globe, as they have recently done, knitting a thin cloak of nervous material around the planet, an electronic “world brain” of some kind is inevitable. The choice for humans is: what kind of internet do we choose to make out of this envelope? Is the system default open, or closed? Is it easy to participate, modify procedures, share, and hide, or is it difficult, burdensome, proprietary? The details of the web can go in a hundred different ways, although the technologies themselves will bias us in certain directions. Yet how we express the inevitable global web is a significant choice we own.

More importantly, the arrival of this inevitable technological stage opens up vast new territories in which we can exercise our free will, despite the inherent inevitabilities in its path. The arc of the technium’s progression contains a clear bias towards increasing free will, options and possibilities.  Technology wants choices.  The internet, to a greater degree than any technology before it, offers choices and options. While the web itself is still embryonic, hardly 6,000 days old as I write this and still a prototype, we can see in this infantile neuronal layer many ways in which this technology can expand the sphere of choices for us personally.

But the technium is expanding not only human choice. It is also extending the long-term expansion of free will in general, non-human and mechanical. This enlargement of volition was first ignited 4 billion years ago by the arrival of life and the birth of tiny things that choose to go here or there, and do this or that. In fact, like the other extropic trends, the increase in free will really began at release of atomic particles in the big bang. As theoretical physicist Freeman Dyson has noted, the exact moment when a subatomic particle decays, or the direction it chooses to spin, must be described as an act of free will. All the microscopic motions of a particle are reversible and predetermined by the laws of physics from the particle’s previous position/state. Yet a particle’s spontaneous dissolution into sub particles and energy rays, or the choice of its direction of rotation are not. Its moment of decay or change in spin is not reversible or predetermined by laws of physics. We call this decay into cosmic rays, or direction of spin, a truly “random” event in an otherwise deterministic realm but this indetermined ” randomness” is really the manifestation of the tiniest quantum bit of free will. Mathematician John Conway, inventor of a computer life-like  display known as the Game of Life, argues that you can’t explain the spin or decay of particles by randomness, nor are they determined, so free will is the only option left. He writes,

Some readers may object to our use of the term “free will” to describe the indeterminism of particle responses. Our provocative ascription of free will to elementary particles is deliberate, since our theorem asserts that if experimenters have a certain freedom, then particles have exactly the same kind of freedom. Indeed, it is natural to suppose that this latter freedom is the ultimate explanation of our own. 

There are other technical arguments for free will in particles. Theoretical biologist and physicist Stuart Kauffman suspects this non-random indeterminism, or free will, is a result quantum decoherence and recoherence of the kind we see in the famous delayed-choice two-slit particle experiment. In this classic demonstration  a single photon is fired towards two parallel slits. But the photon, which is a wave/particle only chooses (note the verb) whether to pass through the slits as either a wave or as a particle after it has already done so and is measured. In the lingo of quantum physics, the decoherence of being both wave/particle (a superposition) is collapsed to a singular choice when the particle is measured later. According to Kauffman the shift in quantum coherence is the source of volition. It’s a wild idea, but the idea that particles have free will is not. 

Very long ago as quantum matter clumped into larger structures such as atomic compounds and spinning clouds of dust and eventual nucleic acids, the tiny slices of quantum volition inherent in particles were leveraged by that increase in organization. For instance, a cosmic ray blasted from a spontaneously decaying particle can trigger a mutation in the highly ordered structure of a DNA molecule. When a “random” cosmic ray knocks a hydrogen atom off of a Cytosine base, say, that indirect volition can birth an innovative protein sequence. Most innovation bring death sooner, but with luck a mutation will confer a survival advantage to the whole organism. Since beneficial traits are retained and built upon by the DNA system, the positive effects of free will can accumulate. Volitional cosmic rays also trigger synapse firings in neurons, which  introduce novelty signals into nerves and brain cells, some of which indirectly nudge an organism to do this or that. By the complex machinery of evolution, these remotely induced “choices” are captured, retained, and amplified as well. Mutations triggered by the free will of particles, in aggregate and over billions of years, evolve organisms with more senses, more limbs, more degrees of freedom. As usual, this is a virtuous self-amplifying circle.


Following the long-arc of evolution, the leading edge of life becomes more complex. The prime way that complexity is revealed is in the increasing ways that an organism can choose. A bacterium has a few choices — perhaps to slide toward food, or divide. A plankton, with more complexity, more cellular machinery, has more options. It can detect and follow more chemical gradients, move toward light, or not. A star fish can wiggle its arms, flee (fast or slow?) or fight a rival, choose a meal, or a mate. A mouse has a million choices to make in its life. Right or left? Now or later? It has a longer list of things it can move (whiskers, eyeballs, eyelids, tail, toes), and a wider range of environments to exert its will upon, as well as a longer duration of life to decide in. More complexity expands the degrees of possible choices.

A mind, of course, is a choice factory, often creating options and demanding decisions that are internally generated (rather than coming from outside). Artificial systems likewise generate a zillion new options, and like bacterium, these systems make unconscious, but real choices. Whenever you send an email, an extremely complex systems of data servers, rules, protocols, and fancy algorithms decide the path of intermediate relays your message will hop along to get to its destination. The path of stepping stones (one out of millions of possible ones) is chosen in real time to minimize congestion, and maximize speed of the network over all. So a second email sent to the same long-distant address even a moment later will require a second choice and is not likely be routed along exactly the same path. The internet is making billions of these non-deterministic free will decisions every day.

A few hours after the big bang 14 billion years ago, the total freedom available within the fine mist of light atoms and zipping particles drifting in the universe was stifling narrow. The possible arrangements between them were dreadfully few. You could count the actionable options for a helium atom on one hand. Compared that prison to the universe one billion years ago (at least in the neighborhood of Earth), when life unleashed an overwhelming explosion of freedoms. Millions of species, each of them an engine of options, filled the surface of a planet with staggering choices. The same restricted hydrogen atoms could now bind with a hundred new elements (created by the stars), in innumerable compounds. Compare that relative cornucopia to today. The technium takes the magnitude of choices unleashed by life and ups it by many exponential orders. We have invented many new ways to arrange chemical elements that do not readily occur in nature, and we have invented new kinds of life, and we have invented machines with new kinds of behavior, never before witnessed in the universe. (Think of hundreds of avatars in a virtual world collectively constructing a treasure hunt.) Once machines unleashed possibilities as fast we could think them up; now they unleash possibilities without waiting for us.

Not only do all these inventions widen the space of what is possible, and stretch the parameters in which decisions can be made, but just as important, the technium creates new mechanisms which can exercise free will. Gadgets such as fuzzy-logic appliances make real choices. Their tiny chip brains weigh competing factors and in a non-deterministic way the fuzzy logic circuits make a decision about when to turn off the dryer, or what temperature to heat the rice. Many kinds of complex adaptive contraptions — for example the sophisticated computerized auto pilot that flew the 747 jet you rode the other day — expand the range of free will by generating new kinds of behaviors out of reach of either humans or other living creatures. An experimental robot at MIT can catch a tennis ball using a brain and arm that is manyfold faster than a human brain/arm combo. This robot shifts so fast while deciding where to put its hand, that our eyes can’t even see it move. Here free will has expanded into a new realm of speed. When you type a keyword into Google it considers approximately a trillion documents before it chooses (and that is the correct word) the page it believes you want. No human can possibly encompass that planetary volume of material. In this way, a search engine gives free choice a scale way beyond the human.

In the world of tomorrow, say a hundred years from now, artificial intelligences and smart stuff will stream self-directed decision-making deeply into the technium. Hi-tech automobiles that park themselves will make as many free-will choices as we do when we park. To varying degrees, technology will practice free will at greater levels than it does today.

New ideas, new technologies contain new freedoms —an expanded range for action. The more powerful a new technology, the greater the new freedoms. This expansion includes possible abuse as well. New technology provides new avenues for freely-chosen horror, as well as good. Present in every new technology  the is the potential to make new mistakes. In fact, unless a powerful technology can be powerfully abused, it is not powerful. Nonetheless, as technology expands so does the space in which our free will operates.

Starting at the big bang, self-organization has steadily increased the range of volition from the tiny choice inherent in elemental particles, to the more visible choices made constantly by living organisms. The self-directed trajectory of evolution continues that expansion into the technium. The technium is designed to expand free will. First by expanding the range of possible choices, and secondly by expanding the range of agents which can make choices.

Although we cannot say that increased technology causes increased freedom, it is clear that multiplying options goes hand in hand with multiplying liberty. Nations of the world with plenty of economic choices, abundant communication options, and high education possibilities, tend to rank highest in available liberty.

“With more choices, we have more opportunities,” declared Emmanuel Mesthene, a technology philosopher at Harvard.  “With more opportunities, we can have more freedom, and with more freedom we can be more human.” 

While every advance in the technium reduces some options (our electronic age provides fewer choices for steam cars), and while evolution converges on invariant forms that may seem to limit human choice, in reality, the technium continues to expand free will as it unrolls into the future. What technology wants is more freedom, expanded free will.

  • Michael Nielsen

    Kevin, you seem to be asserting that quantumness is a key component in free will. At the same time, conventional computers don’t exploit quantum effects in any essential way. It’s true that they do make use of them in certain inessential ways — ideas like semiconductor bandgaps are very helpful in the modern computer industry. But that’s an implementation detail, and it’s possible in principle to build a conventional computer without using quantum effects in any essential way. Danny Hillis’ tinkertoy computer is an example.

    Putting these two ideas together, it suggests that you don’t think the problem of hard AI can be solved — that computers can’t be conscious. Is that right?

    • Kevin Kelly

      @Michael: I think hard AI consciousness is inevitable. We’ll figure it out for sure.

  • Fred Hapgood

    I don’t think that free will, as the term is used in everyday life by human beings, has anything to do with the number of equally accessible states. Free will is about the power to choose, specifically about making the most meaningful choice. A system in which there are only two accessible states — say for instance two paths diverging in a wood — and in which the actor chooses “the least traveled by” for reasons of his own, is just as much the expression of free will as a system presented with two hundred paths.

    Free will, as the term is commonly understood, actually means the opposite of having several choices: it means being free to pick the choice with the most meaning to you, as opposed to being compelled by a billiard ball universe into a choice with less or no meaning. Free will is really just another kind of determinism — the determinism of meaning as opposed to that of physics. It has nothing to do with randomness at all.

  • Michael Nielsen

    Kevin: If one assumes quantum effects are needed for free will, it follows that conventional computers (which don’t need quantum effects) can’t have free will, which means that, presuming we’re using terms in the same way, hard AI consciousness is impossible.

    Since you don’t accept the conclusion of this argument, it seems to me as though you must also reject the premise, i.e., you must think Conway et al are all wrong about the connection between free will and quantumness? Maybe I’m misreading you, or missing some key point.

    • Kevin Kelly

      @Michael: I am agnostic. 1) I think quantum is not “needed” for free will, but is the lowest state we see free will expressed. It may go even deeper than quantum. 2) Free will and consciousness are not equivalent. 3) Who knows what kind of computation we’ll need for AI?

  • Fred Hapgood

    Re consciousness.

    > I think hard AI consciousness is inevitable.
    > We’ll figure it out for sure.

    But how will we ever know? I have no way of knowing that *you* are conscious. How would I ever be able to look at a pile of schematics and say
    circuit X would be conscious and Y would not be?

  • Michael Nielsen

    That clears the argument up, thanks.

    Regarding your point 3), what type of computation we’ll need for AI, I think it’s interesting that the standard model of quantum computing doesn’t qualitatively change what a computer can do, it only changes the speed with which it can do it. Of course, it may be that physics beyond that — say, quantum gravity computers — might expand the range of what computers can do.

  • Per Lindholm

    @Michael Nielsen

    Exactly. If you think of free will as an ideal ability. In reality you would have a simulated version of some sort, then it matters how fast and efficient you can run this free will system.
    Don’t know much about quantum mechanics but nature would find and select for an efficient and fast way of solving this problem.

  • Vodoqc

    Hey, great writing! I was reading the section on technologies that were abandoned -specifically, the ancient Greek device, and it reminded me of an article in this months Wired magazine. It is on the rise and sustainability of mediocre technology – as long as a cheap and easier mediocre product exists, it will be used much wider than a better, but more expensive or less portable/wieldy product. I image that is what happened to these extraordinary Greek devices; they were works of wonders, but the requisite attention to detail and production time made them impractical for consistent use.
    I tried to find a link for you, asit ties nicely with your work here; but no luck.

    By the way, one such technology that should be left in a ditch is spell check. It has done nothing but led to the inability of many people – including myself, it seems, to do basic editing on their own.