The Big Nine

by Amy Webb

  Yet another glitch: an AI crime wave no one saw coming. Narrow but powerful AI programs have started causing trouble all over the internet. They’re making illegal purchases: counterfeit designer handbags, drugs, and medicines made from poached animals (like rhino horn and elephant tusk). They’re listening in on our social channels, reading the news, and infiltrating financial markets by triggering sudden sell-offs. In public spaces, they’re committing libel with the intent of defaming people’s character and reputation. We are beginning to worry about AIs breaking in to our PDRs, hacking our biometrics, and falsifying not just our own records but also those we’ve inherited. Some of this lawlessness was intentionally designed and deployed by the modern mafia: a widespread, distributed network of organized crime that’s difficult to trace and contain. Some of the rogue AIs were accidental: they simply evolved and started behaving in ways no one intended.

  The problems extend to physical robots, too. Security robots, outfitted with smart cameras and predictive analytic software, chase down people of color regularly. The security robots don’t carry weapons, but they do bark very loud orders and sound high-pitched, screeching alarms if they suspect any wrongdoing. Inside of office buildings, hotels, airports, and train stations, people of color are routinely harassed and humiliated because a security bot has mistakenly tagged them as suspicious.

  The G-MAFIA does not have an easy relationship with American law enforcement agencies, which all want access to our PDRs. Rather than working together, the government threatens lawsuits and tries to compel the G-MAFIA to share its data, though it has no obligation—legal or otherwise—to give in to their demands. While no one will go on the record publicly, it’s sounding like US law enforcement agencies hope to emulate some of China’s algorithmic monitoring and social credit score system. Fearing consumer backlash, the G-MAFIA continues to keep its systems locked.

  We had talked for more than a decade about the philosophical and ethical implications of algorithmic decision-making within law enforcement; however, no standards, norms, or regulations were ever established. Now we have a seemingly unending string of AI-powered crimes, but we have no mechanism for punishment. There is no jail for AIs and robots. The laws that define what a crime is don’t apply to the technology we’ve created.

  Our confusion and disillusion has played neatly into the hand of China, which is no longer a near-peer competitor to the United States, but a formidable direct competitor and militaristic pacing threat. China spent decades stealing American equipment design and defense strategies, a tactic that is paying dividends. President Xi is further consolidating the power of China’s military, which is focused on code rather than combat. For example, the beautiful light shows China deployed for various events—a 2017 “drone lanterns” festival, a 2018 summer “drone fireworks” spectacular, for example—turned out to be practice runs for swarm intelligence. China’s military now uses powerful AI-powered drones to hunt in packs all over the countryside and over the oceans.

  Through its economic might, person-to-person diplomacy, and show of military strength, China is practicing a new colonialism, successfully colonizing Zambia, Tanzania, the Democratic Republic of Congo, Kenya, Somalia, Ethiopia, Eritrea, and Sudan. China is building infrastructure—and deploying its social credit score system—and extracting critical resources to lock out competitors and to support its rapidly growing middle class. It now controls more than 75% of the world’s lithium supply, which we need for batteries. And it’s decimated global rosewood forests and led to the extinction of the Mukula tree, a slow-growing species in central Africa that, for a time, was harvested to make red-colored end tables and chairs with intricate carvings.

  No foreign power—not the United States, Japan, South Korea, or the European Union—had enough political or economic clout to stop China from extending its special economic trading zones far out into the South China Sea, East China Sea, and Yellow Sea. Nearly half of all global trade must pass through one of those zones, and every single ship that goes by must pay the Chinese government a hefty tax.

  China observers say that Beijing missed its 2025 target to become the world’s AI powerhouse, even if it has taken control of certain physical world resources. But those observers aren’t looking at the bigger picture. Years of mandatory technology-transfer agreements, uncontained restrictive market practices, plus China’s sizable investment in American and European tech companies proved wildly successful. China now dominates advanced tech industries, including robotics, new energies, genomics, and aviation—and every one of those fields leverage and are leveraged by AI. There are no published numbers, but considering its state AI labs, partnerships with Baidu, Alibaba, and Tencent, and all of its Belt and Road partners, experts believe that China managed to grow the value of its total AI ecosystem more than 500 billion yuan (about $73 billion) in just a decade.

  2049: And Then There Were Five

  As time wore on and progress was made toward artificial general intelligence, the constellation of the Big Nine changed in ways that were both profound and problematic. Now China’s BAT are stronger than ever and still working in lockstep with Beijing. However, America’s six original G-MAFIA members are now only five, due to strategic partnerships and joint ventures: Amazon-Apple and Google-IBM are the four companies that matter most. Microsoft is currently providing support for legacy systems and services.

  Perhaps most surprising is what became of Facebook. It wasn’t the aftermath of Cambridge Analytica or even revelations about Russian meddling into the US elections that led to Facebook’s ultimate demise. Nor was it the fatigue we all felt as our news feeds filled with ever more vitriol, hate, fearmongering, and political conspiracy theories. Facebook’s business model just wasn’t sustainable over time. Once users dropped off and advertisers stopped spending their money on the platform, Facebook didn’t have a diverse portfolio of revenue streams. By 2035, it was in serious financial trouble. Shareholders wanted out, institutional and mutual fund managers got spooked, and the market turned against it. Facebook was sold for parts. Everyone whose data was locked inside the network—which was most of America—is now gravely concerned because it’s our data that was quietly bought by a conglomerate. Investigations are underway, but rumor has it that the conglomerate was, in fact, a Chinese shell company. It’s likely that all of us are now a part of China’s social credit system, and that we’re all being tracked.

  You, like all Americans, are learning to live with constant, low-grade anxiety. Our national sense of uneasiness is often compared to nuclear war threats in the 1960s and 1980s. Except that this time around, we’re not sure what to fear, exactly. You don’t know if your PDR is safeguarded or what personal data China might have access to. You’re unsure of how deeply rooted Chinese government hackers are within our country’s infrastructure systems. You are often awake late at night, wondering what China knows about you, the bridges you take to get to work, the gas lines feeding into your house—and what they’re planning on doing with all that information.

  What we didn’t anticipate was a wide spectrum of AGIs, built for different purposes and tasks, which are both powerful and indifferent to human values. In hindsight, this was remarkably naïve of us. As Amazon, Apple, Google, and IBM partnered, chose sides, and grew, they didn’t set global standards. Decades ago, people bought apps and games for their phones from Google’s Play Store, and because it was fairly easy for anyone to launch and sell an app, the quality varied wildly. There were far too many battery-hungry apps, games that scraped and shared personal information, and janky ads that made the mobile experience miserable. That’s what we’re seeing now in AGIs—except the aftermath is far worse. Some AGIs pretend to follow the protocols written for them but then choose to overwrite those protocols with new directives. Some AGIs self-improve even if their creators didn’t explicitly program them to do so. Some self-replicate, break into other AGIs, and harvest the resources they need to achieve goals, regardless of the impact those actions might
have on the greater ecosystem.

  To counter ill-behaved AGIs, researchers at Applezon and Google-IBM are deploying nanny AGIs—NAGIs for short—to police other systems. NAGIs have a clear set of protocols:

  • To investigate and analyze other AGIs to see if they are violating their original goals.

  • To create a detailed log of all misbehaving AGIs, along with their entire histories (e.g., who created them, when they were modified, and by whom or what).

  • To find the original human in the loop of development, and to notify them of noncompliance.

  • After a grace period (which depends on the severity of the AGI’s infractions), decommission any rogue AGI.

  • Never to modify their own goals.

  It’s evident that both Applezon and Google-IBM were trying to control a system that was starting to spin out of control, but now there is no widespread adoption of NAGIs outside the Applezon and Google-IBM ecosystems. Using previous antitrust rulings against Google and Microsoft as precedent, the European Parliament claimed that NAGIs were nothing more than a hidden attempt by the companies to stifle entrepreneurs and quash competition. The EU became the first to ban NAGIs. Even as research scientists pleaded with regulators to allow these specialized AGIs to help contain what they know is a serious, burgeoning problem, Congress ruled against the tech giants, prohibiting the use of NAGIs in the United States. Those shortsighted NAGI rulings only seeded public distrust in Applezon and Google-IBM, which might otherwise have been good custodians of our PDRs.

  Your home has been turned into a big container for marketing, which is constant and intrusive. You see custom video advertisements anywhere there’s a screen: the smart mirrors in your bathroom and closets, the retractable screens you carry in your pocket, even the smart windowpanes you had to install in your homes to block out extreme solar heat. You are uncomfortable in your own home—the one place you used to feel most at ease and relaxed.

  This distrust has made our health care system particularly daunting. Applezon Health System and Watson-Calico have made tremendous advancements in both AI and medicine. They both got the idea from a mind-controlled robotic suit that debuted at the 2014 World Cup. Duke University neuroscientist Miguel Nicolelis had figured out how to meld mind and machine—and his work inspired others to bring brain-machine interfaces to market. In some tech-forward offices, workers are encouraged wear electronic headbands and link their minds together, along with AGI, to solve challenging problems. Not everyone is comfortable with this high-tech form of collective intelligence since it requires data to pass through either Applezon or Watson-Calico, who can now literally see inside our heads.

  It was Watson-Calico, in partnership with a prominent New York university, that advanced one of Turing’s lesser-known AI theories about morphogenesis. Turing thought that a system of chemicals probably reacted with each other, and that reaction diffused across a series of cells to change some of them. Turing was proven correct. AGI systems were used to discover different ways to create complex multicellular beings, and that led to the advent of augmented human beings, which we refer to as “human-animal chimeras.”

  The original intent was to create viable human tissue for transplants, so we used pigs and sheep to grow harvestable livers, hearts, and kidneys. Researchers also developed brain organoids—the exact same tissue that makes up our own brains. It was promising work, until we realized that AGI was being used to develop human-animal chimeras that had other characteristics, like pigs implanted with human brain tissue that developed a low-level humanlike IQ and newborn babies that had a dog’s sense of smell. What no one has yet discussed (or determined) is the implications of chimera attributes, which are heritable. What happens when a human modified with extrasensory capabilities has a child with someone who also has modifications?

  What’s most concerning is that China decided to repurpose AGI and brain machine interfaces—which were intended to help sick people regain their faculties—for strategic military advantage. It has been used to enhance the cognitive abilities of its soldiers, who do much of their work from inside dark underground bunkers. In the US and EU, such experimentation and use of technology violates ethics laws.

  We are beginning to see a very real decline of Western civilization and our democratic ideals, thanks to China’s colonization, the expansion of its economic zones, and its unscrupulous use of AGI. The health of our economy is in peril, as traditional indicators like housing, construction spending, and food and retail sales are all down, quarter after quarter.

  Even Applezon and Google-IBM are finally seeing a decline in revenue, and they are generally worried about their futures. As they work to overhaul our PDRs to work alongside guardian AGIs, both notice strange noise in the log systems. There are fragments of code that don’t quite make sense, and some of the AGIs that process and route our PDRs are acting glitchy. In a rare act of collaboration, Applezon and Google-IBM share what they’re seeing with each other, in the hopes of determining the problem. In our homes and offices, the lights randomly turn off. Our smart glasses stop working intermittently. Our communications satellites veer off course.

  Though we can’t hear them, we know the shots have been fired, and that China has waged war on America.

  2069: Digitally Occupied States of America

  We realize that China has, in fact, developed a generation of AGIs that have far greater capabilities than ever before seen. Without NAGIs to watch over rogue AGIs, China was able to build and deploy a terrifying system to control most of the population on Earth. If we don’t comply with China’s demands, it cuts off our communications systems. If we fail to keep our data pipeline open to the Chinese Communist Party, it shuts down our critical infrastructure, like our power plants and air traffic control.

  You are a resident in China’s Digitally Occupied States of America. Your transportation, bank, health care system, light switches, and refrigerators are all controllable by China.

  What began as a colonial push into Africa resulted in a new, global Chinese empire enabled and empowered by artificial intelligence. Humanity is on the brink of a terrifying ASI that has been developed by a country that does not share our democratic values and ideals.

  CHAPTER SEVEN

  THE RÉNGŌNG ZHÌNÉNG DYNASTY: THE CATASTROPHIC SCENARIO

  “This is the way the world ends, not with a bang but a whimper.”

  —T. S. ELIOT

  By 2023, we have closed our eyes to artificial intelligence’s developmental track. We missed all the signals, we ignored the warning signs, and we failed to actively plan for the future. We helped the Big Nine compete against itself as we indulged our consumerist desires, buying the latest gadgets and devices, celebrating every new opportunity to record our voices and faces, and submitting to an open pipeline that continually siphoned off our data. We shared silly videos of Alexa failing when our kids chat with Amazon. We asked our TVs to scan our faces, never questioning why a television might need or want our biodata. Every time Google launched fun new projects that map our bodies to photos, our faces to paintings, our voices to celebrities, our fingerprints to people in distant lands, and our irises to our ancestors, we eagerly took part, desperate to keep up with digital influencers and the latest memes.

  AI’s tribes say that diversity matters. It is their mantra. They say it again and again, during keynotes and at conferences, during job interviews and board meetings, in think pieces and tweets. They say it in college brochures. They say it on attractive posters hung in elevators and taped to the hallways at work. AI’s mostly white, mostly male tribes are trained to recite the mantra in their classrooms, labs, and workspaces. Rather than making difficult choices and changes, they stick to the mantra and promise that change is coming soon. And it works just like mantras were intended: to eliminate negativity from the mind and make AI’s tribes feel better about themselves. The gurus in AI’s tribes hand the mantra down to each new cohort of disciples, who feel a sense of great accomplishment in its repetiti
on.

  The mantra echoes within the comfortable bubble of AI’s tribes, which believe they are promoting inclusion when the opposite is true. They champion diversity of all kinds—political parties, religious affiliations, sexual and gender identity, race and ethnicity, economic status, and age—but make no serious effort on inclusion. Rather than seeing a wide, colorful spectrum of people and their worldviews entering the field of AI via tenure track positions, top jobs on research teams, and managerial roles in the G-MAFIA, we instead see no change.