One of the most discussed and sought-after transhumanist goals is immortality. Indeed, the idea of eternal life precedes our postmodern era by millenia. Even when immortality technology was inconceivable, the concept was there. Almost every religion posits that the, raw, distilled “you-ness” will live on or otherwise be preserved after life on Earth. The desire to remain and persist is fundamental and the ultimate cognitive dissonance is considering nonexistence—a dissonance which can be avoided only by ignoring the topic altogether. Or, perhaps, by doing something about it.

“Even when immortality technology was inconceivable, the concept was there.”

Today that manifests itself in increasing numbers of people who expound on potential technological fountains of youth. There are any number of these, from nanoscopic automatic cell repair robots, to enhanced medical technologies making use of stem cells or clones to replace worn out parts, to basic manipulation of the cell at a molecular level to prevent it from breaking down in the first place. The technology I want to examine today, however, is the idea of mind uploading.

I am not a Vitalist in this matter. I don’t believe there is some important, unfathomable part that makes you you, but instead that the brain is a holistic matrix which gives rise to consciousness through enormous numbers of complex neuronic interactions. I wanted to clarify that just for reference, but as I’ll show later on, it really doesn’t matter whether you believe that or not, at least not in regards to uploading the mind for immortality purposes.

Mind Uploading is that idea that, as computers become faster and more complex, there will eventually come a time where we will be able to model an entire virtual brain. Assuming that the only thing required for consciousness is for all the relationships between the parts of the brain and its neurons, and the strength of all the connections, etc, to be in place, when stimulated virtually, this construct should react in a manner indistinguishable from the person’s brain off which it was modeled. I believe in the general principal, but I think we’ll find it’s more complicated than that. How would the brain react without the body, for instance? Because it is tied into the entire body, even if we don’t “think” with our entire bodies. (Or perhaps we do, but that is rather an essay for another day.)

Technically speaking, I can’t think of any reason this shouldn’t eventually be possible—so although it may be impossible, I’m going to presume it can be done. In conjunction with a virtual model of a brain, we could potentially also end up with brains-on-a-chip (where a piece of silicon is created with the same connections and connection parameters), or brains-in-a-vat, or brains-in-a-clone variations, where a version of the silicon (or using a biological technology not yet developed) is placed inside an otherwise functioning body shell.

Again, let’s recap the point of this technology. What we’re talking about here is saving you from dying. If you and your consciousness are the result of a complex perceptual gestalt which seems likely, then any copied representation of the same gestalt would also be conscious just like you. Sounds good, but—

The consciousnesses arising from these two gestalts are not connected.

And this is something of a monkey wrench in the ointment, because it’s going to make the entire process pointless, at least as far as living and persistence and immortality is concerned. You’ll end up with a consciousness and it talks and reacts to stimulus just like you. In fact, it’ll be claiming it is you—and that the process works, it’s amazing, and aren’t you glad you’ve been digitized for all time? But you’ll be standing there looking at the screen saying, “I’m still here; I’m not in there. What happens when I die? I’ve got this great copy, but it is not me.” And the project, while useful for other things, will be deemed a failure as a path towards immortality.

This applies whether or not you are a Vitalist—whether or not you think that fundamental bit or soul exists. If there is something like that, there’s no guarantee it will be copied, in which case the process fails anyway. The idea of copying souls is not a very scientific one, but it may be that this experiment provides the first empirical evidence of a soul. Say, perhaps in the presence of two mind constructs, the soul doesn’t know which one it should return to. I don’t think this will be a problem, but then again if it is, that certainly tells us some information we didn’t know before. If you don’t believe in a soul or other analogous mind-part, then it’s fairly obvious that the copy will not embody you. It’ll embody some other, luckier, you, who will probably end up with a bit of hubris about the fact that he’s not you, because you’re stuck in meat space and will die.

Although this is a basic bit of logic, it needs repeating: a copy of some thing is not that thing. It may be a perfect copy—as copying digital files yields perfect copies—or it may be an accurate copy, even though this still only applies to a certain magnification. But it is still merely congruent.

Copying the brain would most likely yield an accurate copy (it would need to be extremely accurate to be useful), but probably not a perfect copy. We’ve got some error tolerance there because the brain has a lot of redundancy and plasticity. I would assume that a few missing neurons, while not ideal, would not be a major impediment to the fidelity of the copy.

But what’s more interesting is to consider what happens in the computer after the copy is made. I think it’s rather obvious that consciousness won’t be transfered into the machine, but what about the conscious in the machine? Some probably assume that this person (and I think it makes sense to call them a person, body-less though they may be) would live forever because the digital data can just be moved around. Disgusting biological bodies die and decay, but computer equipment just goes obsolete, plus it can be repaired when it breaks down. And if all else fails, back up to a new drive. Looks like these digital people have it made.

Not so fast.

See, every time you move data around on a computer, you’re making a copy of it. (Unless you’re moving it between directories on one drive; then the computer just rewrites the file index, which saves time.) If you try and backup your precious digital brain, you’ve got a different digital brain, with its own consciousness. It’s like the duplicator from Calvin and Hobbes. You get more and more copies, and each one considers themselves the “real” one. There will be no sharing of consciousness in this situation, and—just like in the strip—each one is eventually going to run off and do their own thing. A personality which does not normally put much significance on loyalty is not suddenly going to become loyal to all these other versions of itself just because they share a common source. To each his own.

We don’t have it great out here. Our bodies break down and we die. But the digital AI’s we may eventually have aren’t going to have it much better, because they’re not going to be able to copy themselves, either. Just being in the computer doesn’t change the fact that a copy is not an original, and multiple gestalts will yield multiple consciousnesses.

I’m not even going to into speculation about maintaining the intelligence as the computer does its work. In a neural net, memories are stored concurrently with the processing, but a computer has separate memory, storage, and processing units. If we build a neural net, I believe we’re going to have to integrate these things before we’ll be able to generate a true consciousness, because otherwise we’re dealing with multiple copies of data, each within its own “space”. That could be the blockade to conscious AI. On the other hand, if we simulate consciousness well enough, where’s the distinction? We’d never know, so something like that would necessarily have to be considered conscious. Either way, that’s at least an entire essay on its own.

While it might be the case that we will achieve immortality by medical or biological means, we won’t achieve it with digital means—with one exception that I can think of.

The key, I believe, is preventing disruption of consciousness. This is the grayest of the gray areas in the whole debate, because there are stories of people in comas who recover, or people who get knocked unconscious, and they’re fine. I’ve passed out before. No big deal in the grand scheme of things because I am still here. My memory is intact—except for the parts directly surrounding the passing-out. But I still feel like me. There is also the assumption that this will be the case with cryogenics. If bodies can be frozen and brought back to life, and the brains reactivated, pretty much everyone bets that the freeze-ees will pick up where they left off.

The difference in these scenarios (including the very real, and therefore very important, coma contingency), as opposed to the copying, is that we’re still dealing with a single gestalt. That gestalt has its own consciousness, and so while we may interrupt its processing for a time, it bounces back when given the chance. This is good news.

It’s a fun introduction to philosophy to explain to people how, by their mid-twenties, they’ve had a couple different bodies. The body is constantly refreshing itself, moving out old materials and moving in new ones, so the conglomerate of atoms which makes up your body is always shifting. If I recall, I think the estimate is a complete replacement of every atom in your body every seven years? The specific numbers don’t matter. What’s important is that the arm or eyeball or brain that you have now is not the same arm, eyeball, and brain you had a decade ago.

But you’re still you—so there’s a loophole!

The only way, I believe, to effectively move an individual consciousness from organic flesh to non-organic electrical components is to migrate over individual neurons at a time. If there were some nano-construction technology or other super-precise robot controlled surgery; we could combine this with other “artificial neuron” technology to transplant the brain. But these technologies have to be able to interface with a working organic brain and “invade” (assimilate?) the space, replacing connections with their own versions.

If this becomes a possibility, it solves two problems. Firstly, the “one gestalt, one consciousness” issue would no longer be a concern. You wouldn’t end up with copies of yourself out there, digitally preserved, but alas, not you. (Unless you wanted this for some reason.) Secondly, it still provides the robustness of metal and silicon and plastic, or whatever these things end up being made out of. The great thing about this speculative technology is that, if we can do it once, we can do it a million times, and just continue migrating ourselves forward, the way the body does now. I suspect that eventually even the idea of doing this with components will be quaint or outdated, and we’ll just act right on the organic brain (and body) itself—keeping in mind, of course, that some brains would benefit from the extra strength provided by non-organic materials. Anyone working in dangerous environments (i.e. astronauts) would benefit from a non-organic brain.

It will be an intriguing experiment to run because I believe we’ll reach the point of brain modeling on the computer first, and then we’ll be able to actually test this speculation. If I’m wrong, great, we’re done. Transhumanism has arrived. But I suspect upon completion of a brain-upload test, we’ll discover that the consciousness part doesn’t transfer, and the promise of trans-human immortality will slip just a bit further into the future.

-Ted