By the time I return to visit my hometown in the early summer of 2019, Microsoft has seen an unprecedented resurgence fueled by cloud computing and an ambitious vision for artificial intelligence. Tech is restless, and so is the world it serves. With another presidential election getting underway, a lot of new energy is going into winning rural votes, including proposals to improve everything from USDA programs to rural broadband. While I’m back in Virginia, the New York Times carries a full-page ad from T-Mobile and Sprint proclaiming the time has come to close the digital divide. “Too much of America has been overlooked, underserved, and left out of the digital revolution,” according to the copy that overlays a pastoral photo of farmland that could easily be Campbell County, Virginia. As regulators ponder whether to allow the merger of these telecom companies, a message speaking to rural America seems now to be their emphasis. A few pages later the newspaper focuses on how President Trump’s ban on Huawei Technologies, the Chinese telecom giant, “dashes farmers’ cellular hopes.”
The premise I set out to write about, that technology is going to be more boon than detriment to the rural economies, seems even more relevant today.
On this morning, I speak with David Reid, a delegate in the Virginia state legislature. I learn that these same issues—the divide between rural America and high-tech America—is also playing out in Richmond, where politics are only a little more civil and productive than farther north in Washington, DC. David represents suburban Loudoun County, but was born and raised in rural Buena Vista (pronounced Buoo-na Vista down here) not very far from my hometown. He has a personal calling to rural policy issues because of his birthplace, but he also sees helping rural Virginia as a means of helping urban Virginia. He also knows that with the right infrastructure and skills, anyone can work in tech from anywhere.
“I look at it not just from an altruistic standpoint of view where it’s the right thing to do, but also from a selfish northern Virginia standpoint of view. The more things that we can do to actually help the rural areas of the state be able to attract new businesses to keep people in those areas, it means that there’s less demand for them to want money from northern Virginia to support their efforts.”
David was part of the “blue wave” of Democrats elected to the Virginia General Assembly on the heels of the 2016 Trump election. During his 2017 campaign, David noted that for approximately every dollar that comes out of northern Virginia and goes to the state capital, only about twenty cents comes back. The rest of that money, he said, gets reallocated to other parts of the state that struggle to sustain themselves, subtracting resources from northern Virginia that might otherwise be deployed to improve schools and transit. He sees no way out of America’s broken political system until, geographically speaking, there is more economic equity. So-called flyover communities and states need to “be seen and lifted up.” In other words, urban, high-tech America depends on rural America’s revival as much as rural America depends on the growth and consumer demands that comes from thriving technology-intensive regions.
Today’s workforce can work from anywhere, so long as there is enough broadband connectivity. Ironically, one way to pay for that in rural Virginia is with funds derived from the national tobacco settlement, which prioritizes broadband infrastructure for former tobacco-growing areas like Campbell County. Christopher Ali is an assistant professor at the University of Virginia and has become a leading researcher and thinker on rural broadband. His op-ed for the New York Times in early 2019 called for a national rural broadband plan modeled after President Roosevelt’s Rural Electrification Act of 1936. Over bacon and eggs at the Farm Bell restaurant in Charlottesville, Professor Ali tells me that even though the country has spent billions on rural broadband, it’s a farce because the telecom companies are using copper wires, not fiber, and are erroneously calling slow-speed DSL high-speed broadband.
“DSL is considered ‘good enough’ for rural people,” Ali said.
He sees rural co-ops as an important ally. Rock County, Minnesota, according to Ali, today has 100 percent fiber to houses and businesses thanks in part to bonds issued using credits from wind farms. The Farm Bill reportedly includes money for rural broadband. Ali shrugs, noting that rural communities qualify only if they can show 90 percent of households have no access even to slow speeds. Put simply, if the local provider supplies achingly slow DSL, that is good enough.
But it’s not good enough. And the negative implications are immediately apparent as I sit down with Hunter Bass, the son of my classmate W. B. Bass, who now runs the sod farm with his brother Allan near my mom’s house in Gladys. Hunter reminds me a little of myself at his age. After graduating high school in 2012, he attended university to study computer science. Unlike me, Hunter and his wife have chosen to remain in Gladys—at least for now—near family and friends. They earn a good living, with Hunter working as a software developer at a company in Lynchburg and his wife studying to become a nurse. But their ability to work from home, to start a family and a business, is imperiled by an Internet speed that, at 3 megabits per second, isn’t adequate to fully connect to the opportunities of a digital now, much less those of an AI-powered tomorrow. His parents’ nearby home gets even worse speeds. My aunt, who also lives nearby, is lucky to get 300 kilobits per second.
It’s easy to overlook disparities like this when you’re living in an urban innovation center like Silicon Valley or Seattle or New York City, and especially easy when your job in high tech focuses all your attention on the very frontier of technological possibility. In our conversations and debates about the future of AI technologies, it’s easy to forget that equitable access to technological basics is just as important, perhaps more important, than the next machine learning advancement, the next AI-powered business, or our debates about how we push the technological frontier forward. As technology continues to advance at such an incredible rate, and as business and individuals increasingly depend upon it for their success, the digital divides that we now have will only serve to amplify inequality in the future.
Hunter is optimistic, though. He is pinning his hopes on the dream that his family home in Gladys might one day offer the same infrastructure—the same foundation of opportunity—that he sees in other communities. Just down the road, for example, South Boston, Virginia, hit the jackpot when the Mid-Atlantic Broadband Communities Corporation and Microsoft announced they would team up to build high-speed connectivity and a new tech hub. But we need more than hopes and dreams.
After saying goodbye to Hunter, I stopped by the Governor’s School campus, which was in its final week of the 2019 session. When I was a teenager, my hopes and dreams led to Central Virginia Governor’s School for Science and Technology, which was a launching pad for my career as a computer scientist and technologist, and a foundation of opportunity for me and my classmates. The building, nicer today than when I attended in the 1980s, has three hundred devices connected to excellent broadband for its 136 students. It buzzes with the staff’s excitement for the school’s mission. But despite the great work that the Governor’s School does encouraging the curiosity and creativity of its students, and sending them out into the world with the skills and mind-set they need to be successful, the school remains underfunded. Steve Howard, who teaches programming languages like Python and C#, is training the people who will help create the future of artificial intelligence and the next-generation technologies we have yet to even imagine. In a recent senior class “science scenario,” his students tell him their goal for one assignment is 100 percent efficacy for cancer treatment. Even though it’s unlikely that a group of bright high schoolers is going to cure cancer now, one of them—with the right skills, a lifelong commitment to curiosity and learning, a determination to help others, and optimism—just might someday.
CVGS draws from schools in the region, including Appomattox, Lynchburg, Bedford, and Campbell County. I am curious if anyone from my hometown, about a forty-five-minute drive away, has followed in my footsteps. There is one, a young man named Joshua. He is a valedictorian who will go on to university to study prosthetics. I am moved by Hunter’s and Joshua’s stories. Hunter told me he would have preferred to take computer science before his senior year, but there were no classes. At the Governor’s School, twelve slots went unfilled this year. We can’t afford to let gaps like this persist, particularly in our rural communities, places that need tech jobs and skilled workers to fill them.
It’s been decades since I drove home from school. Even so, the turns and bends in the road from Lynchburg to Gladys feel familiar and bring back old memories. June is such a beautiful month in rural Virginia. The dogwoods, lilacs, mountain laurel, and honeysuckle are blooming. The cloudy skies turn purple and a thunderous downpour pelts the rental car. It feels comforting, rolling hills and glimpses of the Blue Ridge Mountains off in the background. The same houses on both sides of the road have been here since I was a little kid. It all somehow feels reassuring.
On these trips outside Silicon Valley, beyond the reach of the tech industry filter, I’ve been encouraged and even inspired by my old friends the Bass brothers in Gladys, and Hugh E and Sheri Denton in Brookneal. If we convince ourselves that we have a bleak future, then we will take a protectionist and defensive posture. And if we believe that we have a hopeful future, we will build things to make that hope a reality. I saw that with Hunter. Here’s a kid who grew up in the same place I grew up. He went to Gladys Elementary School and then Rustburg High. He saw an opportunity for himself to go do something different, and not in the way where you’re dismissing tradition in order to pursue your own path. He is telling himself, “No, this is my path and I can still respect my family’s roots.” I could hear in his voice a great admiration for what his family has done. He and most of his friends went off to college, which was not the case with my graduating class. But like me, many of Allan’s friends left home to pursue other careers. I think that’s good and bad. Good because we know what the benefits of a college education are. College isn’t the only path to a prosperous future, for sure. There are and should be a multitude of paths for folks to pursue. But we know that a college education brings concrete and tangible economic benefits for folks. The lifetime earning power of someone with a college degree is more than someone with a high school diploma. The bad thing is that Allan’s friends had to go elsewhere because that’s where they saw work and life opportunities. We really need to think hard about how to create more job opportunities for the work of the future here in rural America so people can choose to stay. It’s why I’m so excited about the work that Linc Kroeger is doing to provide digital skills to those entering the workforce and in employment in rural Jefferson, Iowa.
There’s this disinterested, even disdainful attitude that people can sometimes have about those who choose to live in different places, who choose to pursue different paths in life. It’s very easy to surround yourself with the same news sources, the same political views, the same entertainment, the same activities, and the same culture as everyone else around you. With modern technology, with more of our time spent online and on our devices, and with more and more of our connections with one another mediated by social networks, it’s hard to avoid becoming trapped in self-reinforcing filter bubbles and then not to have those bubbles exert their influence on other parts of our lives. Many of my friends and colleagues see those living in rural communities, people who live outside of the urban innovation centers where the economic engines are thrumming right now, in a very different light than I do. That’s not just unfortunate. It’s an impediment to making the American dream real for everyone. The folks I know in rural America are some of the hardest-working, most entrepreneurial, cleverest folks around. They can do anything that they set their minds to, and have the same hopes for their futures and the futures of their families and communities as those of us who live in Silicon Valley and other urban centers all do. They want their careers and their families to flourish, just like everyone else. Where we choose to live shouldn’t become a dividing line, an impediment, to a good job and a promising future.
That’s the American dream. And it’s on all of us to make sure that it works, because in a certain very real sense, if it doesn’t work for all of us, it won’t work for any of us.
See the people. If you have flown from New York City or DC to Atlanta, chances are you’ve flown over my hometown. The next time you fly, open the shade on your cabin window and look down. See the fields and the infrastructure down below. Recognize those are communities. Notice the expanse and its commerce. Someone down there—intelligent, motivated, and ambitious, someone as intelligent, motivated, and ambitious as you—built that and is working hard to make the future better for themselves and those around them.
While this book was in production in November 2019, the Aspen Institute published a definitive study, Rural Development Hubs: Strengthening America’s Rural Innovation Infrastructure. Aspen’s Community Strategies Group painstakingly researched and wrote the report to help policy makers and other decision makers understand rural context and how to build a better rural development ecosystem. The study focuses on the role of a specific set of intermediaries, rural development hubs, that are doing development differently.
In a previous chapter, I proposed the idea of an Apollo space program—a moon shot—for AI. We need that level of unifying ideal. Driving through the back roads of Virginia, we listen to one of my favorite pieces of music, Dvořák’s Symphony No. 9, “From the New World,” which Dvořák composed during his stay in the United States. Neil Armstrong brought a recording of this symphony on Apollo 11. The music is full of optimism, anticipation, tension, and triumph. He was influenced by African American and Native American music, and by the open spaces of Iowa, which he visited at the turn of the last century.
I love classical piano to the point of obsession, and have since I was a little kid. No one in my family listened to classical music. We didn’t have a piano at home. And I never really learned to play. Still, as a kid I was attracted to pianos as these great, complex machines. I was intrigued by the fact that some people could sit and coax it into making something beautiful. I was frustrated, and still am, that there was this gap, a gap that I could never seem to bridge, between making sound at the keyboard versus making music. And as I started to program, I’ve always seen a parallel between the journey of a pianist and that of a computer programmer. The pianist, sitting alone with his or her machine in an attempt to achieve a level of mastery that would enable the making of something meaningful. The programmer, who essentially is attempting the same thing with a computer.
My favorite piano composition is Chopin’s G Minor Ballade, Op. 23, No. 1. I go through phases where it’s the only thing that I listen to, multiple times a day for weeks on end. Sometimes I listen to recordings from different pianists. Sometimes it’s different recordings from the same pianist. And sometimes it’s just the same recording over and over and over. The G Minor Ballade is one of the few pieces of music I can listen to frequently and still have chills run down my spine during certain performances, most often when the pianist hits the double fortissimo at bar 106, which releases all the tension that’s been building up since the double pianissimo in bar 68.
What’s fascinating to me is that some performances, while note-perfect and played by extraordinary pianists, don’t give me those chills, don’t overwhelm me with an emotional response I find hard to describe. I can appreciate less overwhelming performances analytically, but the same raw, visceral emotion just isn’t there. I have no idea why this is, and I doubt I’ll ever be able to figure it out. Which to me feels like one of the great and joyous mysteries of being human.
AI can compose and perform music, even convincingly. But I’ve never felt from a computerized performance anything of what I feel when Murray Perahia plays the G Minor Ballade. I have a very hard time articulating why a Perahia or Vladimir Ashkenazy performance of Op. 23, No. 1, gives me this feeling, while a Vladimir Horowitz or Arthur Rubinstein performance of the same piece doesn’t, even though I would score Horowitz and Rubinstein greater pianists than Perahia and Ashkenazy. I know for sure that other people react completely differently to this music and these performances. A person’s reaction to music is deeply . . . personal. I’ve always thought that the performance of music might be the purest way for one human being to communicate an emotional state to another. I don’t see how an unsupervised machine could effectively create that same emotional connection, either directly or as a broker.
More important, why you would want it to?
Build skills earlier. I am convinced everyone will need the skills to incorporate AI into their jobs, regardless of industry or sector—whether it’s a computer science PhD that will allow you to get a job doing cutting-edge research, college or vocational training in how to build things with the AI infrastructure that others are building, or the skills that we will all need to work with, reason about, and intelligently debate AI-powered products and services as they have an increasing impact on our lives. Students in Iowa, Virginia, and Wyoming all told me they wished that computer classes and training had been available earlier, perhaps even through a 4-H or Future Farmers of America program. We need to provide a foundational level of technical capability to every person on the planet, just like we expect everyone to have some foundational level of math, science, and language skills, so that everyone has standing to participate in the development of AI. This will help to create prosperity for more people, and it will help to ensure that the creation of AI is more diverse and inclusive. The way we create technology is about to change because of AI on two dimensions. Smaller and smaller groups of folks will have more and more power to create highly impactful products and services, and the mechanism we use to make things with technology is going to change from programming computers to do things to teaching computers to do things. The combination of these two things means that technology creation can be more inclusive and diverse than ever, and that if we want the impact of technology to be directed by an inclusive and diverse group of folks, we need to use every tool in the kit to get folks trained in technology.
The magnitude and scope of potential change means that policy makers have a significant role to play. Government, industry, the academy, labor, and organizations promoting the social good must work together across nearly every discipline to develop policy that both encourages the fair development of these technologies, and to help mitigate the worst side effects of their adoption.
There are four overarching objectives that we should seek to accomplish: democratize access to AI technologies and ensure their fair and ethical use; encourage and incent the development of technologies that reduce the cost of human subsistence; help every individual find agency and purpose in an AI-powered future; and defend individuals from misuse of their data, and from harmful uses of AI.
To democratize access to AI technologies, we should consider providing incentives to research labs, universities, and industry to continue to pursue R&D that advances the state of the art in AI, with specific requirements that the technologists producing these technologies package them in ways that are easy for “ordinary” developers to use. We should specifically encourage the development of learning-to-learn and other technologies that seek to lower the barrier to entry to incorporating AI into new products and services. We should encourage students to study computer science, and provide vocational training and education for young people and midcareer folks alike to get training in digital skills for the AI jobs of the future. Moreover, we should make sure that data and AI ethics are a foundational part of both university and vocational training. Every person working with someone else’s data or building systems making decisions that impact human beings should be able to answer the question, “What are my ethical obligations to those whose data I am handling, and whose lives my work impacts?”
Data are the raw materials feeding existing and future AI systems. Folks building AI-powered products and services will need fair access to the data required to train machine learning systems. Encouraging fair, liquid marketplaces for folks to trade data will be extremely important not just for businesses, but for consumers who should be fairly compensated for the data that they contribute to these marketplaces.
The markets are already working to encourage folks to develop technologies that reduce the cost of subsistence, particularly in health care. There is a policy role here as well to encourage the development of systems that reduce the cost of health care while increasing the quality of outcomes, that help reduce the cost of food production and distribution, that decrease the cost of building and maintaining our living and work spaces, and that help to reduce the cost and quality of securing our safety at the individual, municipal, and national levels.
Democratizing access to AI technologies and using technology to better provide for our basic human subsistence needs will both naturally help to provide agency and purpose in the future. Not having to slave away to provide for our basic needs will enable freedoms that humans have never had. AI also has the very real potential to create jobs that we can’t yet imagine, and to unlock human creativity in unprecedented ways. As I write this, I’m envisioning the scene from Star Trek: First Contact where Jean-Luc Picard is walking a character named Lily through a Borg-infested Enterprise. Through the convoluted plot device of time travel, Lily and Jean-Luc are from periods in time separated by hundreds of years, and when Lily asks Jean-Luc about money he says to her, “The economics of the future is somewhat different. You see, money doesn’t exist in the twenty-fourth century. . . . The acquisition of wealth is no longer the driving force in our lives. We work to better ourselves and the rest of humanity.” This may sound like a bunch of horseshit to us twenty-first-century citizens confronted by the reality of current economic systems, but it’s what I aspire to. And if we don’t aspire to things, especially complicated ones, they won’t materialize on their own. We are the stories that we tell.
As was the case in the Industrial Revolution, when a variety of technologies like the steam engine provided a substitute for human labor and fundamentally changed the nature of work, AI will cause disruption. Jobs will go away, and people will need new purposes and new ways to provide for themselves and their families. Accelerating the creation of subsistence technologies, and making sure that the cost-saving gains of those technologies result in lower prices, not wealthier elites, in conjunction with government subsidies and re-skilling, will ensure that the duration and magnitude of disruption are less than they would be otherwise. That said, we must strive to help those whose jobs are disrupted. How do we help them find a new purpose, and manage the personal strain of economic disruption as well as the loss of dignity that comes with losing a job? Government must have a role here, even if it is nothing more than changing the rules of our current markets to incent businesses to help disrupted workers navigate their personal, individual journey through change.
Government must also play a role in providing individuals and businesses with strong data rights and then ensuring that those rights are respected, as well as making sure that individuals, businesses, and governments do not use AI to cause harm to others. This will include developing the capability to detect and defend against AI-powered hacking and human manipulation. And it will include providing legal frameworks for accountability, liability, etc., for AI that interacts with humans.
Provide access to broadband and digital tools for all. If you are too young to remember dial-up Internet access, and maybe never experienced the futility of drip-drip-drip slow connectivity, then surely you’ve experienced these: traffic congestion that makes you late for work for a very important meeting, a slow drain in the kitchen sink when you’re preparing a family meal, a broken chair or door that goes unfixed because you lack the proper tools or skills to fix it. Not having the right tools to build and sustain a business in the digital era is what it feels like for those outside of urban innovation areas. If you live outside of a bustling city, that helpless feeling of not being able to participate in the digital economy can be crippling. You lack access to speedy broadband as well as the digital tools and skills that are practically taken for granted in urban areas. Research comparing US and Chinese Mean Download Speeds by state and region shows real disparities. States such as Mississippi, Wyoming, and Iowa are at the slow end of the chart, while New York and Washington, DC, are at the fast end. China’s slowest regions are still faster than our more rural areas. Reports indicate China’s progress on 5G will eclipse that of the US.
Compensation. One of the thorniest principles to tackle will be how humans are compensated when machines undeniably take on more and more human capabilities and jobs. Some experts have proposed a universal basic income (UBI), which is a noble idea but a poor solution to the problem that it is nominally attempting to solve. That problem—how will people make a living when machines take away their jobs?—could be better solved using other techniques.
For example, my colleagues and I are researching an approach we call “data as labor.” We explained it in early 2018 in a cover story for the Economist. Artificial intelligence is getting better all the time, and stands ready to transform a host of industries. But in order to learn to drive a car or recognize a face, the algorithms that make clever machines tick must usually be trained on massive amounts of data. Internet firms gather these data from users every time they click on a Google search result or speak with Cortana. People “pay” for useful free services by providing firms with the data they crave. Rather than being regarded as capital, data could be treated as labor—and, more specifically, regarded as the personal property of those who generate such information, unless they agree to provide it to firms in exchange for payment.
It requires a profound lack of imagination and no small amount of hubris to believe that the only human beings capable of leveraging AI to build great businesses will be AI experts and elite software engineers, and that the only way forward for everyone else is subsidies from the wealthy elite. If I had to bet, I would put very large amounts of money on the likelihood that AI will create more jobs and prosperity than any technology in history. The sooner that we stop viewing AI as a substitute for human labor and start imagining how it can be a force multiplier for human ingenuity, the better. Imagine a small-business owner with limited resources who will be able to realize her vision by employing AI technologies to give her the equivalent of a massive, supercheap machine workforce to do her “grunt work” and to complement her human workers, who will be freed to create more value than would otherwise be possible.
But let’s assume that it takes a while for this new vision of work, entrepreneurship, and business to materialize, and that folks are going to face economic disruption as their jobs are replaced by machines. Some financial safety net is entirely appropriate and, in fact, necessary. Failing to take care of hard-working members of our community put out of jobs through no fundamental fault of their own is inhuman, inconsistent with faith-based teachings, and un-American. People need to be able to subsist. They need food, clothing, shelter, health care, access to training and education, and the basic means to participate in the workforce. Universal basic income could help provide some of these things. But it is a blunt instrument that fails to address the underlying fact that several of these subsistence needs, specifically shelter, health care, and education, have costs growing faster than inflation or GDP, which means that it will be a struggle for universal basic income to keep up. Universal basic income in the context of free markets also has a ghettoization problem. For example—and perhaps most important, given the negative feedback loop forming around geographic skill-sorting that is resulting in low-skilled workers being displaced from the thriving economies of urban innovation centers—a fixed universal basic income (not indexed to subsistence costs) will result in folks who need UBI the most being pushed to less desirable locations, and potentially completely out of the geographies where, through re-skilling, they might have the most compelling future employment opportunities. If we index universal basic income to the costs of subsistence, then we have a bunch of thorny issues because some of those costs are growing exponentially because of supply-demand imbalances and unequal incomes.
Perhaps a better answer than just plain old universal basic income is UBI in conjunction with a set of technology and policy investments to reduce the cost of subsistence. For instance, it is highly likely that AI and rapid advancements in biotech, like CRISPR-Cas9, will dramatically lower health-care costs while significantly increasing the quality of outcomes. There is a huge focus among start-ups and big businesses to develop AI technologies that can detect and diagnose diseases early when they are easier and much cheaper to treat, and before they can have impacts that reduce quality of life and/or life-span. These new techniques and technologies are already showing tremendous promise. AI, fitness sensors, and ubiquitous mobile computing will be able to lower the cost of routine checkups and diagnostics. And technology may be able to improve the quality of end-of-life care, helping the elderly, mortally ill, and their families make better decisions; improving the agency, dignity, and quality of life of the patient; and reducing costs at the same time. A large percentage of all medical expenses in the US comes in medical care provided in the last twelve months of life. We could spend 100 percent of GDP and the outcome would be no different in the limit. There is no cheating death. There is a role for policy here too, both in paving the way for deployment and adoption of these technologies, as well as ensuring that our insurance marketplaces are functioning as efficiently as possible.
Hope is perhaps one of the most important things that we as human beings possess. The belief that there can be a better tomorrow is the foundation of progress. Hope fuels our resolve to push through the obstacles of the now to reach whatever that future is. And the absence of hope is a bleak thing indeed.
I can’t tell you the number of times after launching an idea, a paper, a piece of art, some software, a complex project, or a company, that I’ve asked myself or those I’ve been lucky enough to work with, “Now what?” When the response comes back, “We’ve done all we can do, and now hope for the best,” the reply that I’ve learned from decades of painful experience is, “Hope is not a strategy.” Hope alone cannot make the future what we want or need it to be.
With the future of technology, particularly advanced automation and AI, there is much to be hopeful about. Hope is monumentally important. But it’s just a down payment on a commitment that we must make to each other to do all the work necessary to redeem our hope with a future that uplifts everyone.