Back-up brains: The era of digital immortality Simon Parkin
How do you want to be remembered? As Simon Parkin discovers, we may eventually be able to preserve our entire minds for generations to come – would you?
A few months before she died, my grandmother made a decision. Bobby, as her friends called her (theirs is a generation of nicknames), was a farmer’s wife who not only survived World War II but also found in it justification for her natural hoarding talent. ‘Waste not, want not’ was a principle she lived by long after England recovered from a war that left it buckled and wasted. So she kept old envelopes and bits of cardboard cereal boxes for note taking and lists. She kept frayed blankets and musty blouses from the 1950s in case she needed material to mend. By extension, she was also a meticulous chronicler. She kept albums of photographs of her family members. She kept the airmail love letters my late grandfather sent her while he travelled the world with the merchant navy in a box. Her home was filled with the debris of her memories.
Yet in the months leading up to her death, the emphasis shifted from hoarding to sharing. Every time I visited my car would fill with stuff: unopened cartons of orange juice, balls of fraying wool, damp, antique books, empty glass jars. All things she needed to rehome now she faced her mortality. The memories too began to move out. She sent faded photographs to her children, grandchildren and friends, as well as letters containing vivid paragraphs detailing some experience or other.
On 9 April, the afternoon before the night she died, she posted a letter to one of her late husband’s old childhood friends. In the envelope she enclosed some photographs of my grandfather and his friend playing as young children. “You must have them,” she wrote to him. It was a demand but also a plea, perhaps, that these things not be lost or forgotten when, a few hours later, she slipped away in her favourite armchair.
The hope that we will be remembered after we are gone is both elemental and universal. The poet Carl Sandburg captured this common feeling in his 1916 poem Troths:
Yellow dust on a bumblebee’s wing,
Grey lights in a woman’s asking eyes,
Red ruins in the changing sunset embers:
I take you and pile high the memories.
Death will break her claws on some I keep.
It is a wishful tribute to the potency of memories. The idea that a memory could prove so enduring that it might grant its holder immortality is a romantic notion that could only be held by a young poet, unbothered by the aches and scars of age.
Nevertheless, while Sandburg’s memories failed to save him, they survived him. Humans have, since the first paintings scratched on cave walls, sought to confound the final vanishing of memory. Oral history, diary, memoir, photography, film and poetry: all tools in humanity’s arsenal in the war against time’s whitewash. Today we bank our memories onto the internet’s enigmatic servers, those humming vaults tucked away in the cooling climate of the far North or South. There’s the Facebook timeline that records our most significant life events, the Instagram account on which we store our likeness, the Gmail inbox that documents our conversations, and the YouTube channel that broadcasts how we move, talk or sing. We collect and curate our memories more thoroughly than ever before, in every case grasping for a certain kind of immortality.
Is it enough? We save what we believe to be important, but what if we miss something crucial? What if some essential context to our words or photographs is lost? How much better it would be to save everything, not only the written thoughts and snapped moments of life, but the entire mind: everything we know and all that we remember, the love affairs and heartbreaks, the moments of victory and of shame, the lies we told and the truths we learned. If you could save your mind like a computer’s hard drive, would you? It’s a question some hope to pose to us soon. They are the engineers working on the technology that will be able create wholesale copies of our minds and memories that live on after we are burned or buried. If they succeed, it promises to have profound, and perhaps unsettling, consequences for the way we live, who we love and how we die.
I keep my grandmother’s letters to me in a folder by my desk. She wrote often and generously. I also have a photograph of her in my kitchen on the wall, and a stack of those antique books, now dried out, still unread. These are the ways in which I remember her and her memories, saved in hard copy. But could I have done more to save her?
San Franciscan Aaron Sunshine’s grandmother also passed away recently. “One thing that struck me is how little of her is left,” the 30-year-old tells me. “It’s just a few possessions. I have an old shirt of hers that I wear around the house. There's her property but that's just faceless money. It has no more personality than any other dollar bill.” Her death inspired Sunshine to sign up with Eterni.me, a web service that seeks to ensure that a person’s memories are preserved after their death online.
It works like this: while you’re alive you grant the service access to your Facebook, Twitter and email accounts, upload photos, geo-location history and even Google Glass recordings of things that you have seen. The data is collected, filtered and analysed before it’s transferred to an AI avatar that tries to emulate your looks and personality. The avatar learns more about you as you interact with it while you’re alive, with the aim of more closely reflecting you as time progresses.
“It’s about creating an interactive legacy, a way to avoid being totally forgotten in the future,” says Marius Ursache, one of Eterni.me’s co-creators. “Your grand-grand-children will use it instead of a search engine or timeline to access information about you – from photos of family events to your thoughts on certain topics to songs you wrote but never published.” For Sunshine, the idea that he might be able to interact with a legacy avatar of his grandmother that reflected her personality and values is comforting. “I dreamt about her last night,” he says. “Right now a dream is the only way I can talk to her. But what if there was a simulation? She would somehow be less gone from my life.”
While Ursache has grand ambitions for the Eterni.me service (“it could be a virtual library of humanity”) the technology is in still its infancy. He estimates that subscribers will need to interact with their avatars for decades for the simulation to become as accurate as possible. He’s already received many messages from terminally ill patients who want to know when the service will be available – whether they can record themselves in this way before they die. “It’s difficult to reply to them, because the technology may take years to build to a level that’s useable and offers real value,” he says. But Sunshine is optimistic. “I have no doubt that someone will be able to create good simulations of people's personalities with the ability to converse satisfactorily,” he says. “It could change our relationship with death, providing some noise where there is only silence. It could create truer memories of a person in the place of the vague stories we have today.”
It could, I suppose. But what if the company one day goes under? As the servers are switched off, the people it homes would die a second death.
As my own grandmother grew older, some of her memories retained their vivid quality; each detail remained resolute and in place. Others became confused: the specifics shifted somehow in each retelling. Eterni.me and other similar services counter the fallibility of human memory; they offer a way to fix the details of a life as time passes. But any simulation is a mere approximation of a person and, as anyone who has owned a Facebook profile knows, the act of recording one’s life on social media is a selective process. Details can be tweaked, emphases can be altered, entire relationships can be erased if it suits one’s current circumstances. We often give, in other words, an unreliable account of ourselves.
What if, rather than simply picking and choosing what we want to capture in digital form, it was possible to record the contents of a mind in their entirety? This work is neither science fiction nor the niche pursuit of unreasonably ambitious scientists. Theoretically, the process would require three key breakthroughs. Scientists must first discover how to preserve, non-destructively, someone's brain upon their death. Then the content of the preserved brain must be analysed and captured. Finally, that capture of the person’s mind must be recreated on a simulated human brain.
First, we must create an artificial human brain on which a back-up of a human’s memories would be able to ‘run’. Work in the area is widespread. MIT runs a course on the emergent science of ‘connectomics’, the work to create a comprehensive map of the connections in a human brain. The US Brain project is working to record brain activity from millions of neurons while the EU Brain project tries to build integrated models from this activity.
Anders Sandberg from the Future of Humanity Institute at Oxford University, who in 2008 wrote a paper titled Whole Brain Emulation: A Roadmap, describes these projects as “stepping stones” towards being able to fully able to emulate the human brain.
“The point of brain emulation is to recreate the function of the original brain: if ‘run’ it will be able to think and act as the original,” he says. Progress has been slow but steady. “We are now able to take small brain tissue samples and map them in 3D. These are at exquisite resolution, but the blocks are just a few microns across. We can run simulations of the size of a mouse brain on supercomputers – but we do not have the total connectivity yet. As methods improve I expect to see automatic conversion of scanned tissue into models that can be run. The different parts exist, but so far there is no pipeline from brains to emulations.”
Investment in the area appears to be forthcoming, however. Google is heavily invested in brain emulation. In December 2012 the company appointed Ray Kurzweil as its director of engineering on the Google Brain project, which aims to mimic aspects of the human brain. Kurzweil, a divisive figure, is something of a figurehead for a community of scientists who believe that it will be possible to create a digital back-up of a human brain within their lifetime. A few months later, the company hired Geoff Hinton, a British computer scientist who is one of the world's leading experts on neural networks, essentially the circuitry of how the human mind thinks and remembers.
Google is not alone, either. In 2011 a Russian entrepreneur, Dmitry Itskov, founded ‘The 2045 Initiative’, named after Kurzweil’s prediction that the year 2045 will mark the point at which we’ll be able to back up our minds to the cloud. While the fruits of all this work are, to date, largely undisclosed, the effort is clear.
Neuroscientist Randal Koene,science director for the 2045 Initiative, is adamant that creating a working replica of a human brain is within reach. “The development of neural prostheses already demonstrate that running functions of the mind is possible,” he says. It’s not hyperbole. Ted Berger, a professor at the University of Southern California’s Center for Neuroengineering has managed to create a working prosthetic of the hippocampus part of the brain. In 2011 a proof-of-concept hippocampal prosthesis was successfully tested in live rats and, in 2012 the prosthetic was successfully tested in non-human primates. Berger and his team intend to test the prosthesis in humans this year, demonstrating that we are already able to recreate some parts of the human brain.
Emulating a human brain is one thing, but creating a digital record of a human’s memories is a different sort of challenge. Sandberg is cynical of whether this simplistic process is viable. “Memories are not neatly stored like files on a computer to create a searchable index,” he says. “Memory consists of networks of associations that are activated when we remember. A brain emulation would require a copy of them all.”
Indeed, humans reconstruct information from multiple parts of the brain in ways that are shaped by our current beliefs and biases, all of which change over time. These conclusions appear at odds with any effort to store memories in the same way that a computer might record data for easy access. It is an idea based on, as one sceptic I spoke to (who wished to remain anonymous) put it, “the wrong and old-fashioned ‘possession’ view of memory”.
There is also the troubling issue of how to extract a person’s memories without destroying the brain in the process. “I am sceptical of the idea that we will be able to accomplish non-destructive scanning,” says Sandberg. “All methods able to scan neural tissue at the required high resolution are invasive, and I suspect this will be very hard to achieve without picking apart the brain.” Nevertheless, the professor believes a searchable, digital upload of a specific individual’s memory could be possible so long as you were able to “run” the simulated brain in its entirety.
“I think there is a good chance that it could work in reality, and that it could happen this century,” he says. “We might need to simulate everything down to the molecular level, in which case the computational demands would simply be too large. It might be that the brain uses hard-to-scan data like quantum states (an idea believed by some physicists but very few neuroscientists), that software cannot be conscious or do intelligence (an idea some philosophers believe but few computer scientists), and so on. I do not think these problems apply, but it remains to be seen if I am right.”
If it could be done, then, what would preserving a human mind mean for the way we live?
Some believe that there could be unanticipated benefits, some of which can make the act of merely extending a person’s life for posterity seem rather plain by comparison. For example, David Wood, chairman of the London Futurists, argues that a digital back-up of a person’s mind could be studied, perhaps providing breakthroughs in understanding the way in which human beings think and remember.
And if a mind could be digitally stored while a person was still alive then, according to neuroscientist Andrew A Vladimirov, it might be possible to perform psychoanalysis using such data. “You could run specially crafted algorithms through your entire life sequence that will help you optimise behavioural strategies,” he says.
Yet there’s also an unusual set of moral and ethical implications to consider, many of which are only just beginning to be revealed. “In the early stages the main ethical issue is simply broken emulations: we might get entities that are suffering in our computers,” says Sandberg. “There are also going to be issues of volunteer selection, especially if scanning is destructive.” Beyond the difficulty of recruiting people who are willing to donate their minds in such a way, there is the more complicated issue of what rights an emulated mind would enjoy. “Emulated people should likely have the same rights as normal people, but securing these would involve legislative change,” says Sandberg. “There might be the need for new kinds of rights too. For example, the right for an emulated human to run in real-time so that they can participate in society.”
Defining the boundaries of a person’s privacy is already a pressing issue for humanity in 2015, where third-party corporations and governments hold more insight into our personal information than ever before. For an emulated mind, privacy and ownership of data becomes yet more complicated. “Emulations are vulnerable and can suffer rather serious breaches of privacy and integrity,” says Sandberg. He adds, in a line that could be lifted from a Philip K Dick novel: “We need to safeguard their rights”. By way of example, he suggests that lawmakers would need to consider whether it should be possible to subpoena memories.
“Ownership of specific memories is where things become complex,” says Koene. “In a memoir you can choose which memories are recorded. But if you don't have the power of which of your memories others can inspect it becomes a rather different question.” Is it a human right to be able to keep secrets?
These largely un-interrogated questions also begin to touch on more fundamental issues of what it means to be human. Would an emulated brain be considered human and, if so, does the humanity exist in the memories or the hardware on which the simulated brain runs? If it's the latter, there’s the question of who owns the hardware: an individual, a corporation or the state? If an uploaded mind requires certain software to run (a hypothetical Google Brain, for example) the ownership of the software license could become contentious.
The knowledge that one’s brain is to be recorded in its entirety might also lead some to behave differently during life. “I think it would have the same effect as knowing your actions will be recorded on camera,” says Sandberg. “In some people this knowledge leads to a tendency to conform to social norms. In others it produces rebelliousness. If one thinks that one will be recreated as a brain emulation then it is equivalent to expecting an extra, post-human life.”
Even if it were possible to digitally record the contents and psychological contours of the human mind, there are undeniably deep and complicated implications. But beyond this, there is the question of whether this is something that any of us truly want. Humans long to preserve their memories (or, in some cases, to forget them) because they remind us of who we are. If our memories are lost we cease to know who we were, what we accomplished, what it all meant. But at the same time, we tweak and alter our memories in order to create the narrative of our lives that fits us at any one time. To have everything recorded with equal weight and importance might not be useful, either to us or to those who follow us.
Where exactly is the true worth of the endeavour? Could it actually be the comforting knowledge for a person that they, to one degree or other, won’t be lost without trace? The survival instinct is common to all life: we eat, we sleep, we fight and, most enduringly, we reproduce. Through our descendants we reach for a form of immortality, a way to live on beyond our physical passing. All parents take part in a grand relay race through time, passing the gene baton on and on through the centuries. Our physical traits – those eyes, that hair, this temperament – endure in some diluted or altered form. So too, perhaps, do our metaphysical attributes (“what will survive of us is love,” as Philip Larkin tentatively put it in his 1956 poem, ‘An Arundel Tomb’). But it is the mere echo of immortality. Nobody lives forever; with death only the fading shadow of our life remains. There are the photographs of us playing as children. There are the antique books we once read. There is the blouse we once wore.
I ask Sunshine why he wants his life to be recorded in this way. “To be honest, I'm not sure,” he says. “The truly beautiful things in my life such as the parties I've thrown, the sex I've had, the friendships I’ve enjoyed. All of these things are too ephemeral to be preserved in any meaningful way. A part of me wants to build monuments to myself. But another part of me wants to disappear completely.” Perhaps that is true of us all: the desire to be remembered, but only the parts of us that we hope will be remembered. The rest can be discarded.
Despite my own grandmother’s careful distribution of her photographs prior to her death, many remained in her house. These eternally smiling, fading unknown faces evidently meant a great deal to her in life but now, without the framing context of her memories, they lost all but the most superficial meaning. In a curious way, they became a burden to those of us left behind.
My father asked my grandmother’s vicar (a kindly man who had been her friend for many years), what he should do with the pictures; to just throw the photographs away seemed somehow flippant and disrespectful. The vicar’s advice was simple. Take each photograph. Look at it carefully. In that moment you honour the person captured. Then you may discard of it and be free.