Rattling around inside a hard drive doesn’t sound like an awful lot of fun — but then, neither does death.
Both eventualities are rather difficult to imagine, but we’ll all have to give them some thought sooner rather than later. Neuroscientist and neuroengineer Randal Koene thinks it’s only going to be another 10 years before we replace parts of the brain with prosthetics. From there, it’s just a matter of replacing each region systematically, to end up with someone whose brain is immortal and electronic. Could the last person to die have already been born?
“My personal bet is yes,” Koene tells me. “Within 75 years we’ll be able to emulate an entire brain. I would very much like to be uploaded eventually. I hope I’ll achieve that. I’m already in my 40s, so we’ll have to get on with it quite quickly. But this isn’t about me. This is something very important that we all need to work on.”
That’s because having your mind uploaded to a computer isn’t necessarily about the quest for eternal life. As a species which produces scientists and explorers, there’s going to be a point where we hit a ceiling of capability. Our fleshy hardware has limitations that will get in the way of our ambition, and, while the brain has been great at allowing our bodies to survive the threats of the savannah through millennia of hominid evolution, it was never designed specifically to master the complexities of global economics, to interface with super smart AIs of our own creation, and certainly not to travel through outer space.
It isn’t about vanity, either, argues Koene: “If we want to keep pace, both with the challenges coming at us and the challenges that we ourselves want to undertake,” he says over the phone, “well, then, we can’t wait for evolution to take place. The only other way to meet future challenges is to evolve ourselves, and to take control of that evolution; and that means we need to be able to access and modify how our minds work.”
Cozied up in modern homes, largely safe from the elements, and all expecting to live to an age when we’re old enough to pass on our genes without any kind of fear of predation, it’s easy to see Koene’s point. The mechanism of natural selection has probably taken our brains as far as it can. It’s reasonable to think that the next stage of evolution is going to be opening ourselves up for a serious tinkering, just as with everything else we’ve discovered. We’ve adapted our environment to suit our needs. Now it’s time to adapt our own biology, too.
“If you take this all the way,” Koene continues, “if you logically follow the route of being able to interface with the brain, through a brain/machine interface, to being able to create replacement parts that can support or enhance the function of a certain brain area — if you keep that going, what you find is that, at the ultimate end, you’re going to be constrained by certain biological limitations. A neuron cannot fire more than [between] 100–1,000 times per second, which is much slower than any transistor can switch. So, as long as you’re constrained to part of your brain being in biology, you’re always going to have to winnow things down to that. But, if you want to get rid of those constraints, and you want access, and you want the ability to modify and tune in to new challenges, then you’re going to have to be able to move what the brain is doing to substrates that evolve along at the right speed.”
Speak to Koene about mind uploading, though, and you’ll see him wince. It’s an imprecise term that conjures images of a digital self moving from one place to another which, most usually, people imagine as some kind of disk space somewhere. That’s not what he’s trying to achieve. Instead, he and his colleagues at the nonprofit organization carboncopies.org are attempting to reverse-engineer a complete working duplicate of an individual’s mind in what they call Whole Brain Emulation (WBE).
The idea is that this emulation could then run pretty much wherever else you like, on whatever suitable system or material you can engineer. That could, indeed, be on some kind of computer space somewhere or it could be in a synthetic brain of sorts sitting in another body of your choosing — biological, robotic, whatever. At that point, you’ll have created a substrate independent mind (SIM), which is Koene’s ultimate goal. Once that exists, an obviously attractive outcome wouldn’t be whole mind uploading, but instead the creation of a set of neuroprostheses which could replace areas of biological brains and replicate the organic tissue’s functions. As medical science gets better and better at extending the use of our bodies, we’re going to need brain function to keep up to make it all worthwhile. Any kind of dementia or trauma to the brain would no longer be an issue, so long as you’re happy to be the recipient of a new wave of mechanical mind parts.
It’s a lovely vision — but is it actually possible?
Well, the first step toward any kind of emulation of mind function is understanding how the human brain works in the first place, and that’s a classic riddle. There is no neuron-for-neuron map of the brain, and even if there was, it wouldn’t necessarily tell us anything about how the thing actually does its job. Nonetheless, the Human Connectome Project, as it’s known, is still a very worthwhile thing, and its aim is indeed to account for every single connection in the brain as well as understand every part’s function. With around 85 billion neurons to deal with — each with possibilities of thousands of synapses — it’s going to take a while. Compare that with the 13-year-long Genome Project that only had the three billion base pairs to sequence, and with no issue of mass interconnectivity.
At present, the best methodology is the 3D mapping pioneered by Kenneth Hayworth, president of the Brain Preservation Foundation, while working at neurobiologist Jeff Lichtman’s Harvard University lab. It took wafer thin slices of a mouse brain and scanned them sequentially with an electron microscope, much in the same way that one might make an animation. The current speed of that process means it takes a few weeks to successfully map a cubic millimeter of gray matter. Given that the average human brain has a volume of 1260cc, it would be a matter of 24,000 years to get the job done. Of course, you could use lots of electron microscope arrays, and new techniques will make things much quicker—but then that only accounts for a complete model of structure. It’s hard to say how long it would take to derive function.
Another way to approach learning how the brain works is the more classic neuroscience method where you zoom right out and try to identify areas of the brain that may be responsible for certain kinds of behavior, but there’s not really enough detail in that alone to be able to make a copy of someone else’s mind. Finally, there’s a school of thought in the middle ground where you take recordings of the brain’s electrical activity and analyze the whens and hows of groups of neurons firing in circuits. That doesn’t inform you of the mapping — let alone a full connectome — but it would help us learn lots of new things about brain function. Koene isn’t concerned where knowledge comes from, and he knows that these scientists don’t have the same far-fetched goals as himself, but each of their achievements offers him, and the wider community, more tools for their task.
“All three of these approaches are useful but, ultimately, the only way you’re going to find out what level you should be looking at, and how to combine these three sources of information, is if you actually try to build something. You use all that data and you make a neuroprosthesis, and it will work to a degree. It works sometimes, and in some ways it doesn’t, and then you learn from those failures and find out more about what you’re supposed to be looking at. So, in my opinion, the only real way to do this is to build pieces, and that’s what I’ve been focusing my attention on at the moment.”
It might sound fairly crackpot, but it’s an approach that’s showing results. In 2011, Theodore Berger from the University of Southern California led a team that managed to do just that — successfully build a prosthetic piece of a rat’s brain. One of the functions of the hippocampal region is to convert short-term memories into long-term ones. Berger et al. recorded the inputs and outputs of this region of the rats during a memory task, and created a chip which would respond with the same signals for every given stimulus. That part of the rats’ brains was then destroyed and, sure enough, they didn’t have the ability to complete the task anymore. Once the neural prosthesis was put back instead, though, they could.
Now Berger and Koene are trying to repeat this kind of success in human subjects as part of a company set up in early 2016 called Kernel. They’re not making any neuroprostheses for people just yet, but they are taking advantage of clinical studies on individuals with epilepsy who have volunteered to undergo procedures involving implants in their brains. For Berger and Koene, it’s just a question of taking measurements of neural activity before and after, but it enables them to test their theories and predictions of what might be going on in a specific region of the cortex. Koene is cagey about what he can tell me at the time of our interview, but promises some interesting content on kernel.co by the beginning of 2017.
“If we can make a neuroprosthesis that can work to some percentage of normal operating functionality, then that’s an incredible learning experience for the next iteration. Or we may even be able to generalize for other parts of the brain, too, and find that now we have a method for making these prostheses in any part of the mind. Making many of these devices eventually becomes the same thing as WBE,” he says.
Outside the realm of neuroscience, though, not everyone agrees that you can treat the human condition as a computer-like tangle of circuitry. British theoretical physicist Roger Penrose and U.S. anesthesiologist Stuart Hameroff, for example, argue that you need to go to a more fundamental level in order to find the brain’s mechanism for creating consciousness.
They have spent years studying structures called microtubules, packed into nerve cell cytoplasm —the places where anaesthetics work to put us under. Their electrons appear to be close enough to show quantum entanglement, so as far as Hamerof and Penrose are concerned an emulation of the brain would be useless unless done at the microtubule level.
Koene rejects this analysis: “If nature was going to go to the trouble of trying to somehow sustain specific arrangements at the quantum mechanical level — which is really, really difficult at the high temperatures in biological tissues, by the way—it’s just not there, but even if what constitutes our minds were encoded in a quantum way, you could still define it mathematically and emulate it eventually.”
How would it feel to have a brain that was part, or all, machine?
“One possible answer is that it wouldn’t feel any different at all,” suggests Keith Wiley, computer scientist and author of A Taxonomy and Metaphysics of Mind-Uploading.
“If the emulation produced a brain that functioned in the same way and all the sensory modality was accurate, then it would feel exactly like your current state. If someone uploaded you in your sleep, you just wouldn’t know the difference when you woke up.”
Indeed, we may even feel better, says Wiley. There are two main scenarios of brain emulation that people generally think of — one is to gradually replace the biological organ between your ears piecemeal, one neuroprosthesis at a time, until all the fleshy stuff has gone. This continuity could offer a more normal-feeling preservation of the self, rather than the other situation where there’s some kind of scan of your brain, a synthetic version is made in one go, the switch flicks, and you wake up elsewhere as your mortal body shuts down. According to Wiley, though, the two are actually identical.
“In both situations, there’s no way that the original brain survives. You can alter the speed, the time, and other facets of either one of the techniques in your thought experiment until they eventually become the same thing anyway. They’re really two ends of the same scale. So, you have to give them the same judgment. You have to decide that mind uploading of any sort either represents survival and preservation of identity or it doesn’t. I think it does.”
Of course, it’s still only a thought-out guess. The answer could just as easily be that we don’t survive. It’s either a slow death or a quick death, but a death all the same, creating something else in your place that is exactly like you and is adamant that it is you, too. There’s little evidence that will ever be able to prove it one way or the other, which is perhaps why Wiley would consider it equal.
He subscribes to the psychological theory of identity, which states that identity doesn’t have anything to do with the physical. It’s not about the brain at all, but our memories and our personalities. In that sense, any subsequent versions of you that are created — even if the original you continues to exist — are equally identified with the same sense of self.
If we agree with Wiley, Koene, and the majority of the mind-uploading community that we might one day live virtually, indefinite lives, then there’s one more fundamental question to consider: How would this change our ideas of humanity, of who we are? What would it mean to be human?
Ernest Becker argued in his Pulitzer Prize-winning work, The Denial of Death, that the knowledge we’re going to die is absolutely central to the human experience. It’s the guiding issue of our lives, and much of what we do is an attempt to distract us from our mortality. We purposefully involve ourselves in so-called “immortality projects” — great, heroic works of one sort or another that we believe will outlast our biological bodies, and so offer some kind of eternal existence. Why would we bother with great works at all, though, if we’ve got no death to fear?
It’s a question forced on psychiatrist Dr. John Wynn. A specialist in therapy for oncology patients and also for the well-being of physicians, Wynn tackles the subject of death every day.
“There are a whole range of outcomes from such a situation,” he tells me. “Skipping quickly over the gross overpopulation of the planet and the inability to sustain the human population if we continued to grow — which would be horrendous — there would also be all this terrible intellectual baggage of all these aging people with their fixed ideas and their ossified concepts dragging progress like a race car pulling a semi trailer.
“The current generation has its ideas, and its pride in its ideas, and every generation has to overcome the fixed notions of the preceding generation in order to generate new knowledge and new techniques. If you’ve ever worked in an academic setting, departments are often controlled by the intellectual apparatus of the senior members. So it’s important for them to die and get out of the way. That not happening could really slow us down as a species.”
Even on an individual level, there’s no guarantee that a longer life would make us any more productive. Serial procrastinators and layabouts would possibly still while away the same percentage of their existence in front of the PlayStation, spending centuries instead of years on the couch. Workaholics, meanwhile, would be beavering away — they may even make copies of themselves, forming some kind of high-powered academic scrum. That said, without that threat of a deadline, they might not be quite so furious in their work. No pressure, no urgency; and so, Wynn agrees, perhaps no progress either.
“Why would reproduction be important at all if the species doesn’t require it in order to go on? Life would be all about accumulation; wealth, power, friendships, whatever people collect — bowling balls, costumes from Africa — but this accumulation depends on resources. We can only take this thought experiment so far. We’re postulating infinite life, infinite resources, infinite space to life. Eventually, there still has to be some kind of check on us somewhere,” Wynn says.
Indeed, neither Wynn nor Koene is particularly keen to talk in the language of forever. Infinites are a very difficult concept to deal with because we can never be sure that something won’t come along and change everything. As such, it’s possible that our brains, running their Darwinian firmware, might never have to, or manage to, feel like they’re going to stick around forever.
Even with the threat of death all but removed, we might not quite believe it. Given that our brains grew around the kernel of survival, of thinking in order to protect from predators, that central tenet of mortal threat could be just too deeply ingrained to remove.
We may not have the brain language to cope with our eternity.