The technology to transfer our brains to computers is coming. But will our digital copies actually be us?
Image: Matt Laskowski/Flickr
Since the dawn of computer science, humans have dreamt of building machines that can carry our memories and preserve our minds after our fleshy bodies decay. Whole brain emulation, or mind uploading, still has the ring of science fiction. And yet, some of the world's leading neuroscientists believe the technology to transfer our brains to computers is not far off.
But if we could upload our minds, should we? Some see uploading as the next chapter in human evolution. Others fear the promise of immortality has been oversold, and that sending our brains off to the cloud without carefully weighing the consequences could be disastrous.
Listen to Motherboard's podcast about brains:
"I think uploading is possible, I just don't think those uploads will be the real us," Susan Schneider, a philosopher who studies the nature of the mind and consciousness, told me over the phone. "This could be good technology for many purposes, but we really need to think about the implications."
Schneider isn't the only one taking this science fictional matter very seriously—indeed, philosophers at Oxford's Future of Humanity Institute recently authored a report on the technological requirements for whole brain emulation. But there may be no one more serious about uploading than Randal Koene, a computational neuroscientist whose life goal is nothing short of transferring himself—and the rest of humanity—to a computer.
"The mind is based on the brain, and the brain, like all biology, is a kind of machinery," Koene told me when I spoke with him over the phone. "Which means we're talking about information processing in a machine."
Like it or not, from a computational perspective, your brain might as well be a larger, more complex version of an insect's
For Koene, the dream of turning this biological machine into an artificial one has always been deeply personal. As a child he felt hampered by the constraints of his own body and mind, and found himself asking how those limitations might be overcome.
"I felt there were so many exploratory things I'd like to do, but I couldn't achieve them all because my brain doesn't work fast enough and I don't live long enough to carry out all the projects," he said. "If you have a computer and want to make a better one, you could figure out how to do so. So I thought, how could you improve the mind? How could I have a better memory, a faster brain?"
The solution was straightforward: Build a new one. A brain that can be augmented and upgraded over time; one unfettered by biological hardware.
"As a species, the fact that we have limitations in our ability to adapt to different challenges is a huge risk going into the long term future," Koene said. "Having a mind that can work in different substrates and change what it's capable of doing is about creating that long term adaptability."
When Koene was in college in the early 90s, the concept of whole brain emulation barely existed. He focused on fields he felt would equip him to tackle the problem, including physics, information theory and electrical engineering. He went on to get a Ph.D in computational neuroscience at McGill University,studying how memories are formed and transferred.
Since completing his studies, Koene's strategy has been to refine a roadmap toward whole brain emulation, moving from one job to the next to push research problems, gather knowledge, and grow a network of like-minded individuals. Following a research position at a neurophysiology lab and a stint as director of neuroengineering at the Fatronik-Tecnalia Institute in Spain, in 2010 Koene landed a job directing research at Halcyon Molecular, a biotech startup backed by PayPal cofounders Peter Thiel and Elon Musk.
"They [Halcyon] wanted to make money doing DNA synthesis and use that to spur growth in artificial intelligence," Koene said. "It was a good place for me, because they encouraged me to put effort into whole brain emulation."
It was during his time at Halcyon that Koene forged a relationship with Ed Boyden, a computational neuroscientistat MIT.
"I invited Ed out to give a talk, and we ended up discussing how you could record activity in the brain," Koene said. "He took the ideas we discussed back to MIT, and formed a group around mapping the physics of brain activity."
That group, whose explicit goal was to map the activity of every single neuron in a mouse brain, expanded beyond MIT's walls. Their work eventually caught the attention of Washington and helped motivate the Obama administration's BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Initiative, a DARPA and NIH-backed effort that supports researchers building tools to map the activity of the 100 billion-odd neurons in the human brain.
Around the same time in 2013, an even more ambitious effort called the Human Brain Project was revving up across the Atlantic. With over 100 partner research institutions and 1 billion Euros in seed funding, this EU-backed program is working to create supercomputer simulations incorporating everything we know about how the brain operates.
"In a sense, the need for functional recording in the brain is now happening," Koene said. "That has led me to refocus my mission."
When Halcyon Molecular quietly shut down in 2012, Koene struck out on his own to form CarbonCopies.org, an organization that serves as a hub for a growing network of mind-uploading advocates. Koene has retooled his own focus, from mapping the human brain's structure and functions, to developing the hardware that'll run our uploaded minds.
At the moment, this entails working with the architects of the CM1K chip—the most advanced in a series of "neuromorphic" chips that mimic the way our brains process information. According to Koene, CM1K, which features over a thousand synthetic neurons operating in parallel, is now being refined and developed for the task of whole brain emulation.
In a sense, all four pillars of the mind-uploading roadmap—mapping the brain's structure and function, creating the software and hardware to emulate it—are now areas of active research. If we take Koene's optimistic view, within a decade, we may have the technological capacity to fully map and emulate a very simple brain—say, that of a Drosophila fruit fly, which contains roughly 100 thousand neurons. And that's when the idea of digitizing the human brain really becomes tangible, because, like it or not, from a computational perspective, your brain might as well be a larger, more complex version of an insect's.
"If you have the tools to get the data from a Drosophila brain, in principle, you can do the same thing for human brain," Koene said. "The neurons and synapses and glial cells are very similar. At that point, it becomes a problem of politics and economics rather than science and technology." Estimates of how much memory it would take to store a human brain vary widely, from roughly one to 1,000 terabytes (it would take roughly 10 terabytes to store the entire printed collection of the Library of Congress).
Which brings us back to the question of whether or not uploading humanity to a chip is actually a good idea. That may depend very much on what we're trying to achieve. If we're simply trying to preserve the knowledge and experience contained within our minds, copying them to the virtual realm seems a logical route. But if the goal is life extension—if we're asking whether a human consciousness can be transferred from brain to computer chip—the answer is far less clear.
That's because, while we've made phenomenal strides mapping the wiring structure of human brain, we still know very, very little about the origin of consciousness, that inner experience that makes us something more than very clever calculators.
"The beauty of our existence is we're not just processing information, we feel," Schneider said. "There could be creatures doing all the information processing we do but that don't have our inner experiences."
And if we embrace uploading technology without understanding what consciousness is, we may, inadvertently, commit digital suicide.
"It could be that, even though it looks to us that silicon is a terrific medium for computation, it nevertheless can't carry consciousness," she said. "I don't think there's any good scientific evidence for this, but it's a possibility we need to consider."
Even if consciousness can be reproduced in silicon (as many neuroscientists believe) there's another, more troubling issue at play here: Whether or not your upload would be a true extension of you.
"I believe that if you have uploaded your brain, whether or not your biological brain still exists, that upload is indeed very much as continuation of yourself."
As Schneider explains, the idea that a person could achieve immortality through uploading is premised on the assumption that the human mind is essentially a program, one that can be copied and run in a different media.
"I think if you coded all of my information, if you got every detail perfect, then yes, you could express an algorithm that follows the same thinking pattern that my brain follows," she said. "But there's a difference between a description of reality and reality itself. If you die, and I make up a new being that's a perfect description of you, it doesn't mean that person is you."
As a thought experiment, we can imagine what would happen if Joe underwent a non-destructive upload; that is, if his brain survived the procedure. Which version of Joe—the original meatbag or the digital copy—would be the real one? Or do they both have equal claim to his identity, his financial assets, his family? You can imagine how such a situation would raise all sorts of thorny ethical and legal questions.
"If the uploading is perfect, these are going to be conscious creatures with rights, and we're going to have to treat them like persons, even if they aren't us," Schneider said.
Koene is less concerned with these personal identity issues. In his view, there's very little scientific support to the notion that consciousness cannot be transferred.
"A lot of people worry about whether there is something very special about consciousness that you do not capture by capturing the functional aspects of the brain," he said. "But we've seen no evidence that there's anything outside of that. Every time we look at a certain function, it depends on hardware in the brain."
Koene also takes a radically different tack to the two-body problem. He suspects our interpretation of consciousness—as a quality bound to a single mind—may be nothing more than perception. Just as our understanding of the physical world is filtered through five senses, or definition of self is intrinsically tied to our limited biological experience. Perhaps we can expand beyond individual minds, evolving into omnipresent, Q-like beings a la Star Trek.
"I believe that if you have uploaded your brain, whether or not your biological brain still exists, that upload is indeed very much as continuation of yourself," he said. "I think a lot of our fears about this stem from the fact that we have a very naïve and simple understanding of how self emerges and how we are here."
If you find questions of consciousness transfer, digital cloning, and jumping into the Continuum somewhat unsettling, take comfort in the knowledge that none of this is actually possible yet. And within our lifetimes, it may not be. But as the digital realm has expanded to saturate our world, we've all found ourselves making personal choices, placing boundaries (or not) around our relationship with technology.
The decision to upload will be no different. Whether we hope to elevate our intelligence, cheat death, or leave behind a virtual copy to continue our earthly projects, the desire to extend beyond our mortal selves is insatiably human. Are the rewards worth the risk? That's a hard choice we all face.
Jacked In is a series about brains and technology. Follow along here.