Chapter 5: Replacing You — The Future of You: Can Your Identity Survive 21st Century Technology? (2021)

Chapter 5: Replacing You. One of the most pored-over definitions of what we mean by 'person' or 'personal identity' must surely be that of 17th-century philosopher John Locke. In his magisterial work, An Essay Concerning Human Understanding, he suggested that the term 'person' be defined as a 'thinking intelligent being that has reason and reflection, and can consider itself as itself. He goes on to suggest a thought experiment, asking his readers to consider the case of the prince and the cobbler. What, he asks, would happen if the soul of the prince left the prince's body and entered the body of the cobbler, taking with it all of its princely thoughts and pushing out the cobbler's soul? Locke concludes it is the prince who survives this transformation. A modern version of this scenario is to imagine two men are in an accident. Let's call them Peter and Paul. Peter's body has been so damaged it has completely shut down, but his brain remains intact. Paul's brain is dead, but his body is miraculously unscathed. Doctors manage to put together Peter's brain and Paul's body. Which person could be said to have survived this transaction? Most of us believe that it is Peter, as his thoughts, memories, and reason continue to exist. They are simply encased in a new physical frame. Of course, our physical bodies do contribute to our sense of identity, which we'll explore more in the next chapter. But most of us tend to agree with Locke that our minds play the truly essential role. What are we to make, then, of the emergence of technology in in the 21st century that can enhance our minds and our intelligence, from artificial intelligence assistants to devices that can interact directly with our brain. If it is our minds that form the core of our personal identity, what impact could this kind of mind-altering technology have on who we are? Take, for example, the world of work. Once upon a time, the job we did formed a fundamental part of our identity. Forged by the physical tasks that we carried out and the labour skills for which we gained a reputation. It was so fundamental, in fact, that we literally took to identifying ourselves by our occupation. Works like Charles Waring Bardsley's Dictionary of English and Welsh Surnames, er, 1901, contain a whole category of such names. The one that caught my attention first was Napier, or 'keeper of the table linen', one of the below-stairs roles required in a traditional aristocratic household. It was my great-grandmother's maiden name and, as is Scottish tradition, it became her daughter's middle name and has been passed down through the family so that it is now one of my own four names. The dictionary gives plenty of other examples too: butcher, butler, carver, chamberlain, cooper, ewer, falconer, farmer, hunter, smith, spencer, woodward— the list goes on. Today, while all of these names remain in circulation, many of their associated occupations have become obsolete. Evidence of how much social and technological change we have seen over the past few centuries. The rate of change has become so much faster that new occupations emerge and then disappear in almost no time at all. For example, it's not so long since there existed the job role of typist. Pools of women would sit desk by desk, bashing away on metal typewriters to turn out paper copies of dictated memos to be signed and then mailed. Today, while we no longer have dedicated typists, we still have typing. I am typing this manuscript now, but much of what I am doing is carried out by a word processing package that sits on my computer. Advances in both hardware and software mean that we no longer require the hard-won skill and expertise of the old-fashioned typist. In fact, there may soon be no need for any of us to know how to type. It's already possible to dictate your text and have it appear directly on your screen, or someone else's for that matter. A recent McKinsey report predicted that between 2016 and 2030, the demand for other types of physical skills, such as operating a vehicle or packaging up products, will fall by 11% overall in the United States and 16% overall in Europe. Technology is therefore having an impact on the attributes we choose to develop, reducing the need for physical skills such as strength and manual dexterity, and the cognitive skills required for repetitive roles such as data inputting or processing. Although people therefore tend to fear that robots are stealing our jobs, the truth is what is happening is not so much a displacement of jobs, but a displacement of tasks. The talents that will become ever more in demand are the ones that automation cannot yet replace: higher cognitive skills such as creativity, originality, critical thinking, opinion forming, ethics, and leadership. These are the attributes that draw an identifying line between human and artificial intelligence, and so are the ones each of us might want to hone and protect. But the truth is that this line is becoming ever more blurred all the time, thanks to the ongoing development of artificial intelligence, design designed to enhance the very same attributes that we would like to place on the human side of the line. In July 2020, OpenAI, the artificial intelligence research facility founded by Elon Musk, Peter Thiel, Greg Brockman, and others, released an application processing interface allowing users to access a new machine learning language model called Generative Pre-trained Transformer 3. Or GPT-3. Trained on vast amounts of text data from across the internet, GPT-3 is designed to perform a task called next word prediction, a task that could see it perform much more complex roles than you might expect. Predicting the next word in a sentence may not sound very useful in a professional context, says Josh Monk, whose company Faculty offers consultancy on artificial intelligence to clients including the UK government. But what OpenAI and others have shown is how powerful this framework can be. It generalises beyond the typical definition of a sentence and is able to answer difficult physics questions, turn legalese into standard English, or perform marketing content creation, he said. Feed some legal text into GPT-3 and it can auto-complete a legal contract. Prompt it with a few ideas and it can turn out the script for a play. One Silicon Valley investor fed in some starter information on how to run effective board meetings, and GPT-3 wrote up a 3-step process on how to recruit board members all by itself. "The iPhone put the world's knowledge into your pocket," says the investor, "but GPT-3 provides 10,000 PhDs that are willing to converse on those topics." As this kind of high-powered AI assistance becomes more prevalent, it could have a huge effect on the way we work. To begin with, this may mean breaking down our current job roles into AI-assisted microtasks so that the AI can apply its machine learning capabilities to train itself in each one. Technology investor Balaji Srinivasan believes that much in the same way that researchers have trained machines to expertly recognise text, So too will professionals such as doctors, lawyers, and photographers be able to pass on their own expert skills. In the medical world, for example, he says, taking skilled information, digitising it, and pulling it in will eventually allow the doctor to just focus on the really hard cases, while auto-diagnosis can take care of other things. Josh Monk says that he can see the potential for what would be a huge step forward in the sophistication of human-to-machine interaction. 'You could imagine fields that involve the synthesis and generation of large amounts of data, such as consulting, auditing, conveyancing, and customer service, being massively upturned by the ability of algorithms to produce 85% ready versions of written documents, contracts, and other correspondence,' he said. And as subsequent versions of GPT-type models become able to deal with far more complex and contextual requests, perhaps we will reach a time where a manager can simply give a relatively complex instruction such update the proposal to include a new section on our data protection procedures, and it would be done. This could also see the emergence of a whole new sector of jobs enabled by and specialized around AI technologies and their application. As Josh Schmunk explained, we may find GPT analysts and GPT specialists who exist purely to help organizations tune and optimize their in-house versions of these models across a variety of use cases. Or we may see organisations or individuals that specialise in producing content, such as marketing, specifically targeted for consumption by these algorithms. So what might this mean for how we think of ourselves and our own potential? Renowned futurist Dr. Ian Pearson says that this technology could soon allow us to, quote, concentrate on the bits we want to do and leave the rest to the machine. By starting with hobbies and bringing them up to professional standards by adding AI capability, we will enable the rise of the polymath, Many people will become highly competent across a range of skills. End quote. He believes that this will have a knock-on effect on the way we work too. He says, people may still have a day job, but also operate on a number of other platforms too. The consequences of this will be that the economy will develop and so will society. People will start more businesses, business turbulence will increase, and poor quality businesses will be wiped out. He said. Of course, there's nothing to say that these AI assistants would just be geared towards work. Perhaps we'll come to use them to maximise our skills, intelligence, and efficiency across every aspect of our lives. We may find we need different AI services to match the different ways we wish to present ourselves in different circumstances. For example, you might want an AI assistant that tonally feels like you at work, one that acts and sounds like you at your most expert and professional, and then another that brings out your more relaxed and sociable side for your more personal time. And why stop there? Maybe you could have yet another AI assistant for something specific, like public speaking, to help you come across as a visionary, proactive, and in a communicative a way as possible. Perhaps you might get all 3 AI services by signing up to a package from the same provider, or maybe you'll find that different tech brands offer a different service each echoing their own brand personality. An Apple service that emphasises your creative qualities, Amazon that encourages your efficiency and productivity, and Facebook that communicates in a way that never fails to capture the most sociable you. Perhaps this is what Sergey Brin, co-founder of Google, meant when he reportedly said, quote, we want Google to be the third half of your brain. Some argue that this is a natural extension of the way that we already project personas in our professional and social lives, a trend that has been accelerated by the advent of social media and its offer of the chance to participate under a pseudonymous cloak of privacy. Balaji Srinivasan explains, to quote, 'In the pseudonymous economy, people earn under one name, they speak under another name, and their real name is yet a third name. Hundreds of millions of people are effectively pseudonymous all day on Reddit, and they just use whatever name suits them. They build reputations under their name, and they can switch to other names. And I think we're going to see a huge version of pseudonymous identities, not just for communication, which we have already had in the West for years, but for earning. Close quote. I think he's onto something important here. As more and more of us become concerned that our own views might not be in tune with those in authority, including our bosses, clients, and customers, I think we will become more reticent to express our own views under our real names. A pseudonymous identity might therefore come in very handy for expressing viewpoints without affecting the identity we use for work, or even for work with one client rather than another. This vision of a multiple identity identity might seem weirdly futuristic, But it's not so different to the way we have always thought about these things. After all, the word person is derived from the Latin persona, which literally means mask, the kind of mask that an actor in Roman times would use in a play. Perhaps we all wear many masks and play many different characters, and the idea of using pseudonymous AI-enabled identities is just a more overt way of organising that. If we do begin to harness these external artificial intelligences, it will open up a world of possibilities for what we can achieve and how our identities might be shaped by such assistance. But it will also raise fundamental questions about the nature of identity itself. For example, where should we draw the boundary between human and machine intelligence? And could a machine ever reach the point where it could be considered as Locke would put it, a thinking intelligent being that has reason and reflection and can consider itself as itself. In short, could a machine ever be considered equivalent to a person? Would such a machine be entitled to the same rights as us? Or would we need a new definition of identity that makes us distinctive from these other beings? These are questions that are already being asked by academics and other experts in the field. For example, in his paper entitled 'Legal Personhood for Artificial Intelligence: Citizenship as the Exception to the Rule', bioethicist Tyler L. Jaynes argues that a set of legal protections would have to be given to all non-biological intelligence systems, to quote, 'insofar as they possess the hardware and software to develop code that surpasses the perceived scope of a human author's initial intent'. Jez puts forward three possible legal protections that all non-biological intelligences could be granted in the future: 1. The right to self-expression, meaning that their observations and opinions would be their own. 2. The right to life, where life means containment on electronic systems and not limited by age. And 3. The right to own their necessary components and any other non-biological components they can acquire. There are further rights listed by Jaynes that follow on from these three basic rights, the most interesting of which is the proposal that all non-biological intelligences would have the right to be recognised as a person before the law. This would include—and here the author gives away his US origins—the right to an attorney, to not be a witness against itself in a court of law, and an indictment before a grand jury. Could we one day find ourselves suing an artificial intelligence, or getting countersued in return? If you think this all sounds absurd, you should talk to David Gunkel, author of The Machine Question: Critical Perspectives on AI, Robots and Ethics, and Robot Rights. Convinced that we must tackle these difficult questions rather than deny that they will ever require our contemplation, Gunkel believes that the idea of AI being given equivalence to a person is perfectly possible, at least in one sense. He argues that a 'person' is merely a recognised legal status determining someone or something as a legal entity in its own right, rather than simply someone else's property. He says, 'In New Zealand, they can extend the legal personality to a river. Does that mean that the river has consciousness? No, it means the river is protected under the same set of rights by which another legal person, whether it's a natural person like you or I, or whether it's an artificial person like a corporation, would also enjoy. The same has happened with animals, and in India, dolphins are persons. I asked David what it would take for an AI to have an identity, not just legal personhood. His answer was that it depends what we mean by identity, which he believes is more open to interpretation than we might think. He said, this question is culturally specific and differs across the globe. In the Western European traditions, identity is considered to be the metaphysical property of an individual who persists in his/her/its essential being despite alterations in what the Scholastics called accidents. The famous version of this is Descartes' cogito ergo sum, the fact that the identity of being a thinking thing persists through changes in the accidental properties of the alterable body. This is not true in other traditions, like Ubuntu in southern Africa. In these traditions, personal identity is not some essential metaphysical property you possess or are born with; it is something that develops out of and in response to the community. So identity, in these traditions, is more of a social and relational concept instead of a metaphysical property. All of this is to say that the identity of an AI may be, as it is for us, a matter of cultural location and specificity. Of course, there are many scholars who believe that even the most advanced technology that we can develop will never be anything more than a tool. It will never be ascribed the rights or responsibilities of a human person. In any legal dispute involving AI, the responsible parties would be whoever programmed, developed, and deployed the AI in the first place. Personally, I find myself falling somewhere between these two camps. I'm persuaded by what Tony J. Prescott, Professor of Cognitive Robotics at the University of Sheffield, calls liminality. The idea that we need to define a new kind of being, neither solely mechanical, nor quite the same as a biological organism. Even this approach leaves all sorts of questions up in the air. If these systems were to deserve an identity of their own, for example, what impact would that have on how we think about our own identity? And if we no longer thought of these beings as machines, could we continue to treat them in the slave-like way that we have always done? Surely we would have to rethink how we lived alongside them. Would they have their own set of identifying credentials? Citizenship? A birth certificate? How would we react if robots took to the streets to demand more rights with the cry of #robotlivesmatter? The age of liminality, of beings that are literally on that threshold between man and machine, is in some ways already upon us. After all, people have been walking around with pacemakers for decades. When it comes to having devices like this implanted in our bodies, we do not seem to discern any effect on our sense of identity. But would the same be true if those devices were designed to interact directly with our brains. Scientists have already developed assistive technology such as bionic eyes that transmit impulses along the optic nerve to the brain to partially replace the vision of those who have lost their sight, and AI-enhanced hearing devices that can selectively filter out unwanted noise in order to focus on a particular sound source. How far do technologies like these need to replace or enhance our natural abilities before they affect who we are. One of the hotly anticipated brain-machine interfaces on the horizon is Elon Musk's Neuralink, which would see a coin-sized implant called a link fitted into a specially drilled hole in your skull and then covered over by your scalp to make it invisible. On its underside, the link would carry more than 1,000 tiny electrodes arranged into threads which would be laced during the surgery into the surface of your brain. The idea is that these electrodes would be capable of reading neurological impulses and sending signals to the brain, as well as wirelessly linking those impulses to external devices. If successful—and there are many technological, ethical, and legal hurdles to overcome—this could allow amputees to control the movement of an artificial limb with their mind, and in the longer term, allow us thought control over external interfaces, so that we could carry out thought-based web searches or messaging as a result. Musk himself has much greater ambitions, believing that the same kind of technology will eventually allow these threads to penetrate further into the brain and access areas relating to memory formation. He has talked of wanting to stream music to our brains and giving us the ability to save and replay memories. Of course, none of these rights now involves consciousness. These are merely smart computers that might be able to simulate some of the connections that we make in the brain. But if every one of the billions of neurons in your brain could one day be connected, it raises the question of whether a device like Neuralink might allow you to back up your memories and ultimately even download them into a new body or a robot body, an idea we will return to in Chapter 7. This kind of extraordinary technology has also long been of interest to the military. For over a decade now, for example, the US Defence Advanced Research Projects Agency has funded a project to develop an artificial, mind-controlled prosthetic arm, with patients in recent clinical trials describing it as 'like having a hand again'. So far, these kinds of projects have involved surgery to implant electrodes, which has limited their use to military or civilian volunteers who have a clinical need. But in 2019, DARPA announced that it was giving grants, rumoured to be around $18 million each, to 6 teams across industry and academia to develop non-surgical forms of brain-machine interface technology that could be of benefit to a wider population. This new initiative, the Next Generation Non-Surgical Neurotechnology Program, or N3 for short, is more ambitious than anything that has gone before. It is hoped that it will eventually enable new approaches to the management of neurological illnesses, but the end goal for DARPA is to create super soldiers, equipped for a new form of telepathic warfare, able, for example, to control F-35 fighter jets with their minds alone. Futurist Ian Pearson explains the way that such technology might work: he told me. If you imagine you are trying to enhance the intelligence of a human being, the way you would do it would be to give them some extra thinking space in the IT and to link that to their brain. The human would then see whatever the problem was and be able to process it at high speed on the computer just by thinking about it. Their mind, or part of their mind, would essentially be running on that piece of hardware. In order to do that, you have to have the hardware running equivalent to the analogue neural network architecture because it has to run the same processes. And if it's going to be conscious, then it needs to be very similar to the way your brain works in a lot of different ways. So once you have that link, you would feel like it was your brain, but bigger. For those who think that this kind of technology may prove impossible, do not underestimate the surprising plasticity of the human brain. In his books, The Brain and Live Wired, David Eagleman explains how the brain is much more influenced by its environment and the information that it gets than we once thought. It is constantly working out how to use this external data to adapt and to learn. He argues that we should therefore view the brain as an extremely fluid system, and that if we can plug in new kinds of data streams, our brains will just work out how to use them. In fact, Eagleman has built a company called NeoSensory to explore this potential and has already released the Buzz wristband, which allows deaf people to, quote, feel sound on their skin. Proof positive of the brain's ability to make sense of the new forms of data, the wristband contains vibratory motors that convert audio streams into sensory patterns on the skin, allowing deaf people to hear sounds and distinguish between them by recognizing the particular rhythm of the vibration. NeoSensory has now built on this system to allow one person's brain to become aware of another person's physiology, such as their heart rate or galvanic skin response. It works by taking the data from your smartwatch and connecting it not just to your Buzz wristband, but to another person's wristband too. Imagine you and your husband are somewhere else in different locations, says Dr. Eagleman. 'You could feel your husband's heartbeat. You'll know if he's feeling stressed, and you might want to call and check in on him,' he said. In the past, we may have wondered what it was like to be in someone else's body. It's early days for this technology, but perhaps we are now starting to get a glimpse of just that, as our intelligence begins to make sense not just of our own thoughts and physiology, but that of others too. The exciting thing is that we do not yet know the limits of the brain's plasticity. If technology such as the Buzz wristband is already possible, why couldn't we train our brain to not only receive information, but send out information to wirelessly control a machine from across the room? Perhaps, in the same way that DARPA hopes to link soldiers to fighter jets, we might too find ourselves able to thought control any kind of everyday machine—from a vacuum cleaner or a smartphone to a more sophisticated workplace robot. Or even robots in outer space. Then things get even smarter once we are able to plug our brains directly into specific devices. Imagine plugging ourselves into some kind of shared server onto which we could extend our mind, offload tasks, store memories, increase our mental capacity, and even connect with the brains of other people connected to the same server. When I asked Haim Pearson to explain how that could work, He responded, let's suppose we get to the point in 2040 or 2045 where direct brain links are starting to appear. They can't upload your entire mind, but they can act as an extension of your memory. So an extension of your intelligence and some of your mind can then run on the cloud. If your mind is running on the server farm and my mind is running on the same farm, there's nothing in principle to stop you and I from exploring the same concepts at exactly the same time because our brains start to overlap at that point. Your brain includes this particular chip, or this particular algorithm, and so does mine. And because we share the same thinking space, we could then share ideas and develop ideas together without directly talking to each other. This raises so many extraordinary questions for us to grapple with. Does this mean we could effectively watch the thoughts of every other person whose mind was relying on the same server farm as us? What about when we're sleeping? Would our extended minds still be working away? And does that mean someone else might be able to access our dreams? And if that is the case, does that mean our dreams will no longer belong to us, but will become as public as any other media is today? Would we be looking at a new era of social media in which newsfeeds are replaced by dreamfeeds? If that ever becomes the case, I'd think we'd all get very familiar very quickly with the privacy settings. It would also be transformative for many industries, especially in creative fields such as advertising, marketing, and design. Those industries love to talk about the creative idea and the collaboration that it takes to originate a big idea and then refine it through a series of drafts and redrafts passed back and forth between designers, writers, account managers, executives, and clients until it's perfected. If we ever reach the era of the shared mind, would we be able to get rid of these interim stages? Perhaps all the people involved wouldn't have to depict their ideas and desired corrections in order to get input from someone else. They could just all think the idea through at once in the Megamind. It would be, as Iain Pearson says, quote, "Like using Google Docs, where more than one person can edit the same document at the same time," but it would be Google Mind Docs. And it could be that two people are interacting, or it could be that two million are interacting, and then what you have is a hive mind. The notion of the hive mind is often associated with a movement called transhumanism, whose supporters are united by the belief that humans can and should transcend their natural biological restrictions using science and technology. We'll return to the topic of transhumanism and biological enhancement in Chapter 6, But its philosophy has interesting ramifications for psychological enhancement too. Some transhumanists believe that if we can plug our individual minds into the kind of shared technologies described above, and so unhook ourselves from the confines of our biology, we may one day be able to live in a bigger, more intelligent, integrated consciousness, a collective consciousness, if you will. Sultan Istoğlu is one of the leading figures of the transhumanist movement in the USA, and he thinks that such a fundamental change to the way we think about our personal and collective identity is not so far over the horizon. He told me, 20 or 30 years into the future, with an implant in your brain, or 40 or 50 years when you can connect with the cloud, you will be able to connect with your spouse and your children directly. It'll be instantaneous, like we get tweets on our phone or direct messages all the time. We're going to be much more integrated with one another. We might not even have an identity. 100 years in the future, I would be very surprised if it's just Zoltán. I think it will be Zoltán and his tribe. My wife, my kids, their children, my grandmother and grandfather. To me, this would be anathema because I cannot accept the idea of not having autonomy over who I am. Any sense of individuality, any sense of identity would have all but disappeared. It's bad enough watching the mob surge and the groupthink on shared social platforms like Twitter, but imagine what it would feel like to have your mind fully caught up in that. Who knows what it will be like, says Zoltán, but he thinks we need to be prepared for a future in which, maybe just 60 years from now, privacy has been consigned to the past. He believes we need to ask ourselves how close we want to be to our loved ones. Do we want to know everything about our partners and our children? He says, I have a 9-year-old daughter and she's just starting to go out and walk herself to school. Does she want me in her mind? These are very challenging questions, and I think with a lot of parents, we're going to have to decide how much we want to be in each other's space. I don't mean physically, I mean mentally. It's a very different space to be in someone's mind. I can't help wondering about the practicalities of all this. What would happen, for example, if you wanted to get divorced? I'm guessing it would be pretty impossible to hide an affair. Would you be able to decide who else gets to be part of your hive mind, or would you need the agreements of the others in your mind tribe? Who would get custody of which memories, or would the memories of all be copied so that everyone leaves with the same mental photograph album of their family history? Could anyone be erased from that? And if you are allowed to leave, where would your mind go? Would it return to the group it was originally linked to, say, your birth family? Moreover, how would you maintain two or more hive minds, one for working with colleagues and one with your family? Presumably the two would need to remain discrete. They would certainly need to be made extremely secure. As hackers who gained access to your personal hive mind group could potentially gain access to the thoughts of your children. As Zoltan says, these are questions that it's worth us all thinking about now. Perhaps the technology we are talking about is decades away, but we would be foolish to underestimate the pace and potential of change and its possible implications. No one would deny that technology has already altered the relationship between our identity and our working lives. A trend that stretches back for centuries. Today, we face far more fundamental challenges to the way we think about who we are, whether it's the development of devices that can expand the capacity of our brains, robots that we can control with our minds, and AI that could justifiably claim an identity of its own. Perhaps even more fundamentally, we need to ask how we can maintain our individuality in a future world in which we become intrinsically connected to each other, where one brain might control several bodies, or one body might have access to multiple brains. In such a scenario, perhaps the very concept of personal identity could not survive.