Transhumanism and the End of Aging

Lifehacker covered a recent article in this issue of New Scientist (this is their death special!) about the transhumanist movement, specifically the use of various techniques to prolong human life--perhaps even indefinitely. Transhumanism, for the uninitiated, is basically a movement of cognitive scientists, biologists, computer scientists, philosophers, futurists, and regular old nutjobs which advocates the use of technology to transcend the inherent limitations of humanity. Depending on the focus of the transhumanist, these technologies can take the form of cognitive enhancements (e.g. implanting computer hardware into the brain to improve memory or cognition), physical improvements (think The Six Million Dollar Man), life extension (we'll get to that), or various other mechanical/biological fusions.

The focus of the New Scientist article is, as I said, on transhumanist solutions to "the aging problem," or (in short) death. If you think you detect a note of condescension in my tone, you're quite right (and quite perceptive)--I don't put much stock in this. My reasons, however, might not be quite what you're expecting. First, though, a bit about the actual solutions discussed.

Transhumanist ideas about combating aging seem to come in two basic flavors, which I'll call the "repair flavor" and the "upload flavor." The repair flavor, in a nutshell, advocates the use of genetic engineering and nano-level technology (such as self-replicating nanobots) to "repair" the damage done to the human body over time and keep it running indefinitely. Aubrey de Grey, University of Cambridge biotechnologist and beard enthusiast, compares such maintenance to the sort of preventative work an owner of a classic car does to keep it in working order, saying

"In the same way that we have 100 year old cars now that are working just as well as they did when they rolled off the production line ... similarly, once we become able to implement sufficiently comprehensive technologies, we'll be able to do the same for the human body [...] The people keeping these vintage cars on the road aren't doing anything more sophisticated now than they were 50 years ago, or even 30 years ago, when they were only two or three times as old as they were designed to be. Once you learn how to do sufficiently comprehensive maintenance, that is it. You keep the machine at a manageable level of damage, so to speak--a level that is not prejudicial to the functioning of the machine."

This is an apt analogy, I think. The human body is, after all, nothing more (or less) than an organic machine produced by naturalistic (rather than intentional) processes. There's even some evidence to support the idea that this approach might be viable--Cynthia Kenyon, a biochemist at the University of California, San Francisco, seems to have developed a technique that can prolong the life of tiny worms called Caenorhabditis elegans by up to six times, the equivalent of giving a human a 400 year lifespan. While Dr. Kenyon's technique is rather simple (it seems to work by just inhibiting a single gene), and therefore unlikely to be as successful in humans (whose genomes are significantly more complex), it is a first step in the right direction. Thus, the idea that we might one day visit the hospital for gene therapy, inject tiny robots into our bloodstream to repair damaged cells, and live to see the turn of several centuries might actually be viable. Or it might not.

I have a deep and abiding mistrust for futurism--the path that future technology will take is notoriously hard to predict, and this is particularly true for bleeding edge technologies (like nanotechnology and bioengineering). If someone predicted the internet in its current form 20 years ago, we'd be impressed, but the groundwork was already well in place by 1987--certainly it would have been easier to foresee the internet as it is today from 1987 than to foresee the course of (say) quantum mechanics from 1800; the shift toward quantum mechanics (and away from Newtonian physics) represented such a new scientific paradigm, that foreseeing it would be nothing short of oracular. It seems to me that we're at a similar stage in our development of nanotechnology--while maybe the groundwork is already there, we're nowhere near as close to putting it into production as we were to putting together the modern internet in 1987.

The second path proposed by the transhumanists to overcome mortality is (of course) the upload flavor--that is, uploading our minds into computers. Oxford philosopher (!) Anders Sandberg explains:

"My personal view is that we are the information, the processing, going on in our minds, so an uploaded copy of me would be me [...] a rough sketch would look something like this. You freeze the brain, you slice it very thinly. These very thin slices you scan using an electron microscope ... you get these images, you process them in a computer to create a three-dimensional model of where are the different synapses and neurons, what are the strengths, what are the connections, use that to create a simulation of the brain, then you start the simulation."

If you've talked to me about philosophy for more than thirty seconds, I bet you can guess how I'm going to respond to this. No. Just, no. Might this technique work? Sure, given sufficient complexity and precision of measurement, we could probably set up a perfect model of the brain. With enough computing power, we could probably even run a functional simulation of the brain's activity, but is that enough for immortality? Definitely not. Think about what Dr. Sandberg is saying: create a simulation of my mind, and that's just as good as my mind--even more, it is my mind. Why is this? What's the error here?

The problem lies in Dr. Sandberg's acceptance of the computational theory of the mind, which (unfortunately for him) is probably wrong. CTM basically states that, as Dr. Sandberg says, what makes a mind is the information processing going on inside it. CTM is, at its heart, a functionalist view, asserting that it doesn't matter how the information processing is done, only that the same information is being processed somehow. That is, it doesn't matter if the data from my senses is being processed by my biological brain or by a simulation of my brain--the result should be the same.

This is, as I said, almost certainly wrong. I'm not going to belabor this point too much here, because showing why it is wrong is one of my paramount philosophical projects, and I've already covered it several times here. Suffice it to say that an informational processing model of mind doesn't account at all for the subjective nature of mental states--while a computer simulation of my mind might be able to (with perfect accuracy) predict what I would do, say, or even think in any given situation, it would emphatically not be thinking anything. Thought (i.e. the subjective experience, the "what it is like" factor to be conscious) is something that is forever unattainable to digital computers, at least given our current model of computing. Thoughts, experiences, sensations, and other qualia require semantic content rather than just symbolic content--see Searle's Chinese Room for evidence of this.

Eripsa has engaged me about this argument before, arguing that we shouldn't tie our definition of mind to qualia and the subjective nature of mental states--he says that this is dangerously anthropocentric, and that computer "mind" might take other forms. He might be right, and I've been willing to concede this point, but that kind of argument won't work here: we're not trying to create artificial minds with this technique, but rather to prolong the existence of already functioning minds--our own. For this to be successful, the technique needs to replicate all aspects of our minds, including (and especially) consciousness. We need more than information processing, because we're more than information processing.

Summary: transhumanism would be awesome, but it's unlikely that it will take the form these people think. Genetic engineering and nanotech might give something like this someday, but trying to predict how and when at this juncture is virtually impossible. Uploading our minds to computers is virtually impossible given our current model of computing (unless you're content with a simulation of yourself), and though this gap might be (and probably will be) bridged someday, it is undoubtedly much further off than the repair flavor of anti-aging technology.