The Big Nine

by Amy Webb

  As the tribes’ worldviews become increasingly myopic, the problems we’re already seeing compound. Accidents and mistakes are on the rise, like computer vision systems misidentifying people of color and blaming them for crimes. Surveillance expands while simultaneously becoming less obvious. The line between our personal data and the data we generate at work blurs, and so do the criteria for who gets to use our data and when. Transparency into AI systems fades into darkness. (Not that it was great to start with.)

  The G-MAFIA are the sole owners of your personal data record, which grows to encompass every aspect of your human existence: what you write in emails, the texts you send to your kids, your digital breadcrumbs as you search for the perfect desk chair, the unique contours of your fingerprints and face, where you walk and the pace of your runs, who you bump into at the grocery store, whether you have the flu, and what medications you’re on. Algorithms make decisions for you using all that data. They determine whether you get a discount when booking a flight. They help or prevent you from getting a job, qualify you to buy a home or a car, match you for first dates, and tell the doctor whether you’re lying about how much you drink, smoke, and exercise. Since it’s Google, Amazon, Apple, Facebook, Microsoft, and IBM who own that data—and because we love their products, even if we don’t entirely trust the companies—we can’t see total corporate control of our PDRs for what it is: America’s version of China’s social credit score system.

  We find ourselves locked into a digital caste system, where AI makes choices and judgments based not only on how we’ve lived our lives but also on the PDRs of our parents and relatives. Wealth no longer matters. Status is determined by “being our best selves,” where “best” got defined long ago by a relative few programmers who thought an organic ketogenic diet, midday yoga classes, and regular trips to the chiropractor were the keys to an optimized existence. If you don’t take a weekly infrared sauna, the AI system you’re tethered to will record noncompliance in your PDR. And that act of rebellion doesn’t just affect you, because your record is linked to everyone you know and are related to. You cannot escape the sins of your associates.

  In the near future, Amazon and IBM will persuade the governments of the United States, United Kingdom, Germany, and Japan to open up access to a trickle of citizen health data. Apple, Google, Microsoft, and Facebook will have a more difficult time in Europe because of previous antitrust lawsuits. But those early Amazon and IBM experiments will prove useful to government agencies, which will open up more lucrative contracts for the entire G-MAFIA.

  Back in 2008, when parts of the world entered a financial crisis caused by the housing bubble, China was glad to buy iron, oil, and copper from Latin American countries, effectively protecting those countries from serious harm. When oil prices dropped in 2011, China was willing to invest in and bail out Latin America.1 In 2013, China launched joint military training exercises off the Brazilian coastline—and did it again in 2014 off the coast of Chile.2 In 2015, China’s Defense Ministry hosted a 10-day summit on military logistics with officials from 11 Latin American countries, and in the years since it has invited Latin American military officers to career development programs in China.3 While the American government is retrenching and retreating from the world’s stage, China is in expansionist mode. It is working deals all across Southeast Asia and Africa—and Latin America, too.

  After a decade of steady relationship building throughout Latin America, today it is China—and not the United States—supplying Venezuela, Bolivia, Peru, and Argentina with Chinese military equipment, which includes aircraft and arms.4 And it has a reason to establish bases all across America’s backyard. In Patagonia, China built a military antenna and a space control station, and it built a satellite-tracking hub in northwest Argentina.5 All of this activity involves artificial intelligence.

  Now, policymakers and lawmakers alike are failing to make the connection between China, the US, and AI. China’s consolidation of power under Xi Jinping, its various state-sponsored initiatives, its rapidly growing economy, and the success of the BAT are an unstoppable—if invisible—force with which to be reckoned. Neither the White House nor Congress see that China’s push into all these countries—Tanzania, Vietnam, Argentina, and Bolivia, for example—has to do with both economics and intelligence. They refuse to acknowledge that China is building a 21st-century empire on the foundation of data, AI infrastructure, geopolitics, and the global economy. It is a grave error in judgment we will all later regret.

  Chinese citizens are learning to live with automated monitoring and consequences of stepping out of line. Crime is down, and social unrest is curtailed, and for a time the middle and upper classes preserve the status quo. They have access to luxury clothing and handbags, designer furniture and statement cars never imagined by their parents and grandparents. Promises are made to lift all Chinese people out of poverty. For now at least, it seems like privacy, religious freedom, sexual identity, and free speech are reasonable trade-offs for earning a desirable social credit score.

  US government leaders don’t take enough time to get educated on what AI is, what it isn’t, and why it matters. Aside from the usual conversations about AI disrupting productivity and jobs, those in DC make zero effort to engage the G-MAFIA in serious discussions about other pressing issues related to AI, such as national security, geopolitical balance, the risks and opportunities posed by artificial general intelligence, or the intersection of AI in other fields (such as genomics, agriculture, and education).

  With no strategic direction on AI from the White House—and, in fact, an openly hostile stance on science and technology—Washington focuses on what matters during the next election cycle and what will play well on the Sunday morning political shows.

  Neither the G-MAFIA nor their executive leadership are intentionally putting democracy in harm’s way. But safeguarding America as the dominant global superpower and ensuring the preservation of democratic ideals just isn’t central to their corporate values. Beginning in the early 2010s, Google’s former chairman Eric Schmidt worked admirably and tirelessly to boost US military and government preparedness in the era of AI. It wasn’t a ploy to win government contracts for Google. Schmidt was concerned about our national security and military preparedness in this new technological age. But it was such an unusual undertaking that Silicon Valley questioned his motives. Rather than other G-MAFIA leaders following his lead, they were skeptical of his ambitions. And so aside from Schmidt, none of the G-MAFIA’s leadership have given much thought to the role AI is playing in the rise of China as a possible superpower successor to America.

  There is no strategic collaboration between the G-MAFIA and government agencies or military offices—at least not without a lucrative contract. The G-MAFIA agree to the arcane, outdated procurement requirement policies of the military and government, but this doesn’t accelerate AI in our national interest. If anything, it shines a bright light on the cultural differences between Silicon Valley and DC, and it slows down modernization. The few government agencies built for innovation—the US Digital Service, the US Army’s Futures Command, the Defense Innovation Board, and the Defense Innovation Unit Experimental (DIUx) initiatives—are brittle in their youth and subject to defunding and staff reductions as the revolving door of political appointees spins. Washington views its relationship with the G-MAFIA as transactional. Neither lawmakers nor the White House makes an honest effort to develop the kind of relationships with G-MAFIA executives necessary for a long-term coalition on AI. The G-MAFIA, US military, and government circle around each other without ever converging in our national interest.

  We allow ego and habit to get in the way of building consensus on China. Government officials, trade representatives, journalists, technologists, and academics debate China, the United States, and AI ad nauseam, holding tight to their longstanding, cherished beliefs without making room for alternate realities. The usual suspects argue that Xi Jinping won’t be in power long, even with ter
m limits abolished. Once he’s gone, all of China’s long-term AI plans will evaporate. Their usual detractors argue back: Xi will unite his people and party. Regardless of whether he dies young or cedes his post to a successor, the CCP will be stronger as a result and will see the AI plans to the end. And so it goes, back and forth: China’s industrial policies will have zero impact—or they will cause the unraveling of the US economy. China’s military poses an existential threat to the Western world—or it’s just an overblown trendy story that we’ll be bored with soon. We ought to invest the time and money on a national AI strategy knowing that China’s plans could fail—or we’re wiser to save our time and money and take a wait-and-see approach. There is one point everyone seems to agree on: if America truly gets into trouble, the G-MAFIA will be compelled to help us out.

  Our policymakers, elected officials, and think tanks make the same, tired arguments but take no action. They settle into stasis. They settle into stasis because in America it is difficult to escape the centripetal force of profit without a powerful intervention.

  We have heard the story of stasis told many times before. We preserved the status quo of cigarette smoking, debating hard data about cancer while continuing to market tobacco as an accessory of fashionable women, a pick-me-up for factory workers, and a medicinal remedy for people who were sick. We failed to act on climate change, arguing over and over about adjectives. If there’s global warming, why is it so cold? We resigned ourselves to debating timeframes. The alarming claims made in the 1970s became dire in the 1990s and then apocalyptic in the 2010s, but we’re all still here. Who’s to say things will be all that bad in the future?!

  Systematic change has a compounding effect and builds over decades, not days. By the time we realize that stasis was the wrong course of action, it’s too late.

  2029: Digitally Locked In and Out

  For the past decade, you’ve been incentivized to buy all manner of smart technologies and AI systems. All appliances now come standard with AI systems. Your refrigerator tracks the food inside. Washing machines—even those at Laundromats—track the progress of your dirty clothes, pinging you once a cycle has completed. Your oven shuts off before the turkey burns and dries out.

  But there’s a catch you didn’t see coming: you lack permissions to override what was supposed to be a “helpful” AI. After you put bags of lunchmeats and cheeses, trays of cupcakes, and six-packs of beer into your connected refrigerator—all bought for a Super Bowl party—a notation is made on your PDR. The number of servings and calories exceeds the number of people in your household, so the AI concludes you are planning to overeat. It may be after midnight, and you might have already planned to get a load of laundry done in time to put it in the drier before work the next morning, but the washing machine’s AI doesn’t take your desire to sleep into consideration. It sounds an alarm and pings you—repeatedly and without ever stopping—to tell you that it’s time to put your clothes in the drier. You’d like to make your own turkey jerky from scratch, but the oven won’t allow it, because its AI has been programmed with the goal of juicy meat, period. (Or, if you can afford it, you can pay to unlock the jerky upgrade.)

  Some households experience AI glitches, especially with their kitchen appliances, and typically in the morning. The control panels will go dark intermittently, which unfortunately locks down the door and prevents you from taking out breakfast. The dishwasher will suddenly stop midcycle, keeping glasses and silverware soaking in soapy, greasy water. The volume of smart speakers will suddenly spike, too, making it impossible to talk to your family members over cereal and coffee. You, along with tens of thousands of consumers report outages, and each time the G-MAFIA dedicates a few product managers to research what’s going wrong. Tech journalists attribute the glitches to the “spooky ways” in which “AI acts weird sometimes.”

  At first, the attacks seem novel and random. So we all blame Google, Apple, and Amazon for faulty products and crappy customer service. Then cybersecurity experts are gobsmacked to discover all the glitches are actually linked. It is a new kind of “Internet of Things” attack originating in China and enabled by machine learning. The Chinese have a name for it: , or bèi kùn, which translates to “trapped.” The hackers, backed by the Chinese government, thought it was clever to launch “bacon” attacks during breakfast hours in America and to effectively trap our food, drinks, and eating utensils in our AI-powered appliances. Their purpose is singular and sophisticated: to seed mistrust in the G-MAFIA.

  Microsoft and IBM are still around, but they are minor players in the AI space. Microsoft, which at one point published industry-leading research on computer vision, machine reading comprehension, and natural language processing, never successfully gained internal alignment and momentum on how to compete in AI. Now the company is downsizing and primarily supplying support to its legacy systems: what’s left of its original Azure cloud, SharePoint, Skype, and Outlook. While IBM’s Watson found partners and clients, IBM’s cloud service, which had long been a distant third to Amazon and Microsoft, shrunk once Google began offering competitive rates for both government and big corporations. Its other business units—such as data centers, storage, and semiconductors—have found it impossible to compete against companies in Taiwan, which are now the world’s largest suppliers. For Taiwanese companies, the CCP’s “One China Principle” translates to significant market advantage, even if Beijing restricts their personal liberties and freedoms. China’s industrial policy has effectively prevented IBM from doing business in many areas of the world.

  As for Facebook? After years of promises to shore up its security and provide better transparency into how it shared our data, the majority of its original users have moved on to other platforms. Gen Alpha kids (the children of Millennials) may have had their photos strewn all over Facebook, but they themselves never created accounts. Facebook is quietly going the way of MySpace.

  With interoperability still a critical weak point in the West’s AI ecosystem, by 2035 we settle into a de facto system of segregation. Our devices are hooked into Google, Apple, or Amazon, and so we tend to buy only the products and services offered by one of those three companies. Because the data in our heritable PDRs are owned and managed by one of those companies—companies that also sold us all the AI-powered stuff in our homes—we are Google families, Apple families, or Amazon families. That designation comes with unintended biases.

  Apple households tend to be wealthier and older. They can afford all of Apple’s sleek, beautiful hardware products available in one of three colors: palladium silvery-white, osmium grey, or dark onyx. Apple’s smart glasses, smart toilets, and custom refrigerators carry on its long tradition of pricey products anyone can use right out of the box. Apple’s PDRs come with spoken interfaces and a choice of two soothing voices, Joost (who has a “unisex higher tone”) or Deva (who has a “unisex lower tone”). But convenience comes with a cost. Apple’s AIs cannot be overwritten. In an Apple home running the air conditioner, you can’t open the door for more than a minute or the system will start beeping incessantly. If there’s sufficient daylight detected by the sensors in your light bulbs, the Apple system keeps the light switch on lock-down.

  We saw a preview of Google’s connected home decades ago at the 2018 South By Southwest Festival in Austin Texas. Back then, the tagline was “Make Google do it,” and attractive spokesmodels took small groups around the three-story home to interact with AI-powered appliance screens and connected frozen daiquiri makers. Google’s system is less intuitive, but it makes better use of our PDRs—and it offers different levels of service and access. For those who can afford the upgrade fees and have enough tech savvy, Google Green gives families the ability to manually unlock their systems, and they can connect a greater variety of things—such as coffee makers, 3D printers, and outdoor irrigation systems—to their homes. Green families can also opt out of marketing and advertising, though their data is still collected and sent to third parties. Google Blue is a
n affordable option with limited unlocking privileges and some additional permissions, but Blue families are still subjected to marketing. Google Yellow is the lowest tier. It’s free but comes with no override abilities, has a small selection of available devices and appliances, and has limited data protections.

  Amazon went in an interesting, but ultimately lucrative, direction. A few announcements Amazon made in the fall of 2018 went largely unnoticed, such as the launch of its Amazon Basics microwave, which includes a voice interface. Users could put a bag of popcorn in the microwave and ask Alexa to pop it. Tech journalists wrote the microwave off as a novel, silly use for Alexa and missed the bigger picture: the system was actually designed to get us hooked on subscription popcorn. That’s because the microwave tracks both what we’re heating up and what we’re ordering on the Amazon platform. A new box arrives before you ever have the chance to run out.

  Because Amazon was smartest in its approach, working with federal, state, and local governments—offering them deep discounts at Amazon.com, patiently working through procurement requirements, and building and maintaining cloud services specifically for them—it became the preferred platform for certain social services in the United States. That is how Amazon discovered how to leverage the long tail of government funding.

  Low-income families now live in Amazon Housing, which has replaced city-funded public housing programs in the United States. By every measure, they are far superior to any public housing ever provided through our previous government programs. Amazon Homes are completely outfitted with connected devices in every room. The former Supplemental Nutrition Assistance Program (previously known as the Food Stamp Program) is currently hosted by Amazon, which provides steeply discounted Amazon-branded household products, food and drink, toiletries, and books. Unsurprisingly, this program works seamlessly. There are never delays in funds being distributed, it’s easy to look up the status of an account, and all transactions can be completed without ever having to wait in a long line at a government office. Those living in Amazon Homes must buy most of their things through Amazon while their data is scraped, productized, and monetized for various initiatives. Amazon’s AIs are the most pervasive, following Amazon families everywhere they go to collect valuable behavioral data.