A Silicon Curtain is Descending: Technological Perils of the Next 30 Years
Editor's Note: This piece is part of a full report, "Reassessing 1989," which looks at the major events of that year, including the fall of the Berlin Wall, the Tiananmen Square protests, and the breakup of Yugoslavia.
It could be said that the principal story of 1989 in Europe is a story about technology – of radio and information crossing the East-West divide to bring down the Berlin Wall. Indeed, the post-communist narrative became that more connectivity and more connection meant more freedom and more democracy. It was on the wave of this narrative that the Internet became the world’s ultimate connector.1 It has brought globalization and international commerce in an unprecedented and unimaginable way, given activists a platform and a megaphone, and made information about democratic governance available to anyone with a router. Or almost anyone.
Not half a year before that fateful fall day in Berlin, the Chinese Communist Party (CCP) had sent tanks and troops with assault rifles into Tiananmen Square to suppress student-led, pro-democracy protests in Beijing (and throughout China). This has become one of the most censored events in modern history. When the Internet entered China in the 1990s, the seeds of control entered with it. Subsequent decades have seen the Great Firewall ensure that information about the massacre, as well as information damaging to the party, or to “stability,” is inaccessible in mainland China.
As the China case indicates, the post-communist narrative that connectivity implies freedom has not been airtight. At the same time that innovation brought tools of openness, it enabled further means of control. Looking forward, the complex relationship between technological innovation and freedom that has characterized the past 30 years will only grow more complex in the next 30.
By 2025, the world’s totality of data is expected to reach 175 zettabytes (10 raised to the 21st power). Devises connected and producing data will have an online interaction every 18 seconds. How we manage, store, and derive value from that information will determine national economic and military competitiveness. And emerging technologies that harness this data revolution will define the 21st century relationship between freedom and innovation. Depending on how states choose to use them, they risk redrawing old lines around geopolitics.
Artificial intelligence, quantum computing, biotechnology, and the very infrastructure of the Internet are four of the immense technological revolutions that will shape and are already shaping the globe. Ultimately, all are fonts for endless economic and social possibilities that can shape our worlds for good. They also have the potential to be exploited by autocratic regimes to advance repression and control, sometimes at the same time. Indeed, the seeds of this use have already been planted. And aspects of the data revolution – sensitivities to personal privacy and government accountability chief among them – may handicap liberal democracies and strengthen the authoritarian model.
Democracies have an opportunity now to steer 21st century technology in the direction of freedom by understanding their disadvantages in the data age and working to counter them. The solution lies in keeping democracies competitive and bringing clarity on ethical frameworks.
Hailed as the technology of our time, artificial intelligence (AI) allows us to turn a cornucopia of aggregated data into useful and lucrative insights about the world. AI – and more specifically machine learning – have the potential to transform a myriad industries: healthcare, transportation logistics, telecommunications, automotive, advanced electronics, and many more. According to McKinsey’s “Notes from the AI Frontier,” AI will create trillions of dollars of economic value. Social good applications span education, urban development, ocean life protection, traffic safety, media bias, carbon sequestration, transparency in governance, energy, and nutrition, among others.2 According to Russian President Vladimir Putin, “whoever becomes the leader in [AI] will become the ruler of the world.”3
The benefits of many of these cases will distribute equally to democracies and authoritarian states alike. In some cases, the relatively transparent governance and robust, bottom-up innovation that liberal systems provide may even be necessary to realize societal gains. But there are reasons that democratic governments should not take for granted their continued economic and technological prominence in the age of AI.
Human data labelers working long hours on little pay can produce labeled content on a mass scale.
First, liberal democracies have a data disadvantage. Artificial intelligence systems at their root are classifiers – distinguishing road signs from trees, people from cars, etc. As such, they rely on massive quantities of data to “learn” one class from another. In the case of an autonomous driving system perceiving the road, for example, knowing whether a certain frame or image from a camera on the vehicle contains a stoplight or not requires seeing many images with and without stoplights in the training stage. In fact, providing this labeled training data is exactly what we humans do when we encounter CAPTCHA systems asking us to prove our humanity by clicking on the images that contain stoplights or cars. Similarly, in identifying individuals in a facial-recognition system, the more training images of a person the system has, the more readily it will recognize him or her. In amassing these datasets, illiberal states without strong privacy frameworks may have an advantage. Additionally, because many AI algorithms need labeled data (for example, “this image does not contain a stoplight”; “this one does”; “this is Mr. Smith”), regimes such as China may build labeling factories that would be inconceivable in liberal nations with stronger labor protections and standards. Much like today’s factories, human data labelers working long hours on little pay can produce labeled content on a mass scale.4
Second, liberal states suffer from an explainability handicap when it comes to implementing machine learning systems across society. AI systems make recommendations and decisions based on reams of aforementioned studied data. However, exactly how those decisions are arrived at remains a mystery, even to the engineers coding the AI. Put concretely, if an AI system considers a hundred factors in determining whether to grant a loan, and decides to decline the loan request, by and large it will be unable to generate an explanation as to why the loan was declined. Whereas human decision-makers can create pro/ con lists and decision rationales, state-of-the-art AI systems cannot. For autocrats interested in making the best decision without a populace or strong legal system that can hold the government accountable, that may work just fine. For societies that champion equitability, fairness, and transparency – upheld by a court of law – AI’s explainability issue poses problems for widespread implementation. When AI systems have already shown to propagate existing societal biases in gender and race, such transparency is all the more important. Autocrats do not have these handicaps.
Facial Recognition and Societal Surveillance
In practice, authoritarian regimes are already using AI for suppression and control of populations and political narratives. Deep-learningpowered facial recognition software tracks China’s ethnic Uyghur population through a ubiquitous network of cameras in China’s Xinjiang region. Often described as an epicenter for the application of emerging technologies for authoritarian control, Xinjiang has seen over one million Uyghurs put into concentration camps for trivial offenses such as having contact with relatives outside China, growing a beard, and attending a mosque. Concerningly, these documented human rights abuses have been enabled in part by the technological diffusion of globalization. Indeed, in some cases, Western tech firms have wittingly or unwittingly lent expertise, credibility, or technology itself to building the Chinese surveillance state.
The surveillance systems enabling this frightful control are not contained within Xinjiang. In a 2017 show of force, China’s network of over 100 million cameras was able to track down a BBC reporter in Guiyang, a capital city of about 3.5 million in southwestern China, within seven minutes. Furthermore, China is exporting its surveillance technology around the globe. Zimbabwe, Malaysia, the Philippines, Ecuador, the Gulf, and others have signed up for Chinese city-surveillance packages.5 Russia too plans to expand its own facialrecognition pilot project to 105,000 cameras in Moscow.6 With this export of surveillance technology comes training on how to use it and the authoritarian worldview in tow. Missing in action in many cases are pro-liberal privacy and human rights frameworks to go with the AIpowered surveillance packages.
Beyond facial recognition, the applications of AI for surveillance and control are equally alarming. The same AI-based speech recognition software that may enable near-simultaneous language translation in the near future can also enable simultaneous “public opinion monitoring.”7 In Xinjiang, Uyghurs’ online activity is monitored; throughout China and its user base around the world, technology, likely fueled by AI, censors dissent on WeChat. In some cases, individuals have been jailed for online comments. Far from an age of freedom, the authoritarian Internet is one of control.
5G and Undersea Cables
The future Internet and the backbone for an estimated 50 billion connected devices by 2020 will also be influenced by who controls its infrastructure. Here too, the technological predominance of the U.S.- led liberal coalition is not assured. Future 5G networks will power the full spectrum of the Internet of things – from autonomous vehicles and smart homes to advanced manufacturing plants and electrical grids.
How democracies choose to structure these networks now will have geopolitical reverberations for the next 30 years or more. Questions about the control of next-generation connectivity have surfaced most prominently in the global debate over Chinese telecom giant Huawei’s embedding in worldwide 5G networks. Europe is taking center stage in this struggle. Based on U.S. intelligence community findings that a mammoth Chinese enterprise with an unclear and nontransparent relationship to the CCP represents an unacceptable national security risk in future networks, the United States refuses to allow Huawei components in its 5G plans. And it is urging allies in NATO and the Five Eyes intelligence-sharing alliance to do the same. But Huawei has already made substantial inroads in Europe and around the globe that make extracting it nearly infeasible economically.
As an analyst at a cyber threat intelligence firm recently told The Guardian, “The breadth of technologies and range of information that Huawei could have access to…will likely be too great an opportunity for Chinese intelligence and security services to pass up.”8 Beyond the strict information security risks of backdoors to suck out our Internet traffic and the data of our connected lives, the bigger question is what happens if a Chinese-controlled company controls the world’s entire Internet. Concentrating power and market share in the hands of an authoritarian-based global behemoth will surrender our future Internet backbone to its control, including the ability to shut down parts at will.
China's investments mirror Russia’s inroads in Europe with oil pipelines in projects such as Nord Stream 2 – and we have seen how this infrastructure influence can play out.
China is also cementing its Internet and communications infrastructure control with the placement of undersea data-carrying cables beyond the Asia-Pacific. Chinese state-owned telecom providers China Unicom, China Telecom, and China Mobile are owners of the new SeaMeWe-5 cable connecting Europe, the Middle East, and Southeast Asia. China Unicom also partially owns a cable connecting Cameroon and Brazil. And Huawei Marine Systems – a joint venture between Huawei and British company Global Marine Systems – is building such cables throughout Africa.9 These investments mirror Russia’s inroads in Europe with oil pipelines in projects such as Nord Stream 2 – and we have seen how this infrastructure influence can play out.
The geopolitical significance of Internet infrastructure is illustrated by the case of Vietnam, where Chinese investors have dominance in physical and digital infrastructure. When Vietnam criticized China’s stance vis-à-vis the South China Sea, Chinese investors froze energy infrastructure projects in Vietnam. And when in 2016 the Permanent Court of Arbitration rejected China’s territorial claims in the South China Sea, a Chinese hacker intruded into screens and sound systems in Vietnamese airports at Hanoi and Ho Chi Minh City. The hacked screens broadcasted propaganda messages criticizing Vietnam’s territorial claims in the South China Sea (which conflict with China’s). With their systems down, staff at the airports had to check passengers in manually for several hours.10 A similar stunt at Heathrow or Charles De Gaulle would have drastic economic consequences.
In light of the vulnerabilities of an authoritarian controlled Internet backbone on the one hand, and the Internet censorship authoritarians deem necessary for governance control on the other, Alphabet CEO Eric Schmidt and others have predicted a bifurcated Internet along ideological lines. Indeed, Russia has embraced Huawei’s 5G solution and has already called for its own Internet. The logical conclusion to this course is that a new “silicon curtain” of digital connectivity threatens to replace the Iron curtain that lifted 30 years ago.
Quantum computing, poised to be the next fundamental revolution in computation, has transformative technological, economic, and geopolitical implications for how we process and secure information.
This technology harnesses the properties of quantum physics – the laws of the universe that govern the behavior of electrons and particles in matter – to solve a new class of computational problems and achieve processing times impossible for even the world’s fastest supercomputer. Even as it opens novel societal applications, a full quantum computer has the potential to render vulnerable our most secure personal, commercial, and even military communications. Much in the way that the leaders in 5G technology will set its standards and ultimately control its use, the geopolitical upshot of the quantum-computing race will be that its victors dictate the future of secured information and reap the benefits of processing it.
Three overarching posited applications of quantum computation are especially salient for our global digital future. First, and most challenging to realize, quantum computing holds the possibility to break modern encryption and upend the way we secure information. Second, and related, quantum physics can also be harnessed for an encryption technique called quantum key distribution. This offers a way to shore up communications in a post-quantum world when current encryption techniques are broken. And third, and most immediate in the short term, quantum computing can boost data processing speeds and help solve the computational processing challenge of AI algorithms on the massive data sets of the future; it can thereby improve AI and optimize it for our connected future.
For these reasons and others, the United States and China are investing heavily in quantum computing research and development. The winner of that race will gain significant informational advantages and may ultimately hold the cards in the AI era of amassing, safely storing, and processing data.
Biotechnology in particular will see rapid advancement from a proliferation of genetic and health data. Actors who own that data can drive medical advancement and cure disease, but also employ genetic information for surveillance and the development of sophisticated bioweapons.
By 2025, 40 percent of the datasphere will be in health – the largest of any sector or industry. At the same time, the cost to sequence the human genome has dropped precipitously, from nearly $100 million in 2001 to under $1,000 today.11 The explosion of genetic and health data – and increasing abilities to process it – hold tremendous potential for scientific and medical achievement worldwide.
The future of personalized medicine offers researchers and drug developers the ability to target therapeutics to an individual’s precise genetic makeup. Research is already underway in the United States and China into personalized (and potentially far more effective) treatments for diseases including cancer, cystic fibrosis, and Alzheimer’s. CRISPR gene-editing technology has renewed the promise of genetic engineering with applications such as more nutritious crops, fighting genetic diseases, developing new antibiotics and antivirals, and even the much-hyped (and much criticized) possibility of “designer babies.” In law enforcement, we have already seen DNA databases from commercial genetics companies generate crime suspects, solve cold cases, and even put the long-sought-after Golden State Killer behind bars.
The United States’ position as the global biotech leader is not assured into the next 30 years.
But the United States’ position as the global biotech leader is not assured into the next 30 years. China last year unveiled a $60 billion yuan ($9.2 billion) 15-year research initiative in precision medicine. Further, through research partnerships, investments, mergers, and acquisitions, China has engaged in a systemic exfiltration of biodata from the United States. This data will be the fuel for many next generation applications.
Much as in applications of AI writ large, authoritarian regimes may benefit from fewer privacy scruples in collecting and using biodata for national advancement. In March, Russian President Vladimir Putin decreed that all Russians would be assigned “genetic passports” by 2025. Compulsory “free health checks” in China suck up individual health information. Whereas personal health information in countries with strong privacy protections is considered some of the most sensitive, autocrats can collect and use it largely at will. Even worse, in the case of U.S. biodata, there are legal question marks as to whether the Health Insurance Portability and Accountability Act protects the health information of U.S. citizens when it is transferred overseas.
Global leadership in biotechnology is not solely a matter of economic competitiveness and national wealth, though those elements are important to secure authorship of the rules of the global technological order. It also has implications for the moral and ethical frameworks of these technologies.
The same lack of meaningful public scrutiny that advantages authoritarian regimes in data collection has already found its way into testing practices. In November of 2018, a Chinese researcher announced he had delivered two babies genetically modified to be resistant to HIV using CRISPR gene editing techniques. The announcement was met with an outpouring of public criticism, including at least nominally from the CCP for its reckless humantesting practices. In June 2019, Russian molecular biologist Denis Rebirkov told Nature he was thinking about implanting his own geneedited embryos by the end of the year.12 If researchers in China and Russia discard ethical and precautionary measures around modifying the human genome in ways the rules-based liberal international order will not condone, how can democracies and their moral frameworks remain state-of-the-art and the gold standard in genetic technology?
Even more concerning than how autocracies can use data and ethics advantages to outpace the United States in biotech is how the CCP and other autocratic regimes can misuse it. Xinjiang, the epicenter of the Chinese surveillance state, has received attention for its frightening network of facial recognition-enabled cameras that produce a nearconstant eye on the ethnic Uyghur population. What is less discussed is how genetic surveillance is a part of that picture, enabling authorities to target individuals precisely by genetic makeup and ethnicity. The national security implications of next-generation bioweapons are even worse. Targeted viruses or bioweapons that could wipe out an entire population, all individuals with a certain genetic marker (or all individuals who have not been implanted with a certain marker) are not outside the realm of possibility in a future war.
The next 30 years of exploding data will revolutionize biotechnology, often aided by factors such as lax restrictions on privacy and the rule of law. Democracies need to think outside the box and recognize these global trends to stay competitive and secure moving forward.
Winning Others to Our Side of the Curtain
Today’s moment is perhaps closer to 1989 than to 1946, when Winston Churchill introduced an audience in Missouri, and the world, to the Iron Curtain. By 1946 the Soviet Union had already secured its control over Eastern Europe; in 1989 the future was thrown wide open. Western democracies today do not yet find themselves on the smaller side of a silicon curtain, having lost the technological, economic, and ideological battle. Nor are the disadvantages outlined here meant to imply they will. But they can. To succeed, democracies must marry moral frameworks with strong technological achievement in three ways.
First, we can join with likeminded nations in recognizing and countering democracy’s disadvantages in the data age. A strong transatlantic relationship is as vital today as it was in 1946, and there is rebuilding to do.
Second, we can invest jointly in technical offset solutions to blunt authoritarian advantages. Novel research in “privacy-preserving” machine learning and “explainable AI” models that attack weaknesses in data aggregation and democratic accountability are sound places to start. In the medium-term, a more critical look at harnessing data’s economic power while preserving democratic values is needed.
Last, we must establish and champion moral and ethical frameworks and standards around new technologies that accord with liberal values in a renewed commitment to human rights and the rule of law around the globe. This action is especially important where authoritarian technology is diffusing rapidly and the rule of law receives waning traction.
The real danger is not that liberalism will necessarily lose the technological battle wholesale. Rather, the four revolutions discussed risk diminishing liberal power to win over those teetering states – not democracies but not quite authoritarian satellites either – because we can no longer pair the liberal moral framework with superior economic achievement. It was this economic superiority that characterized the post-1989 era and, before it, ultimately did bring down the Wall – at least as much as did Berlin’s airwaves. Only by countering our techno-economic weaknesses, investing together in the solution, and championing the morals that unite us as integral facets of our global offering can liberalism hope to realize some of 1989’s promises of openness and connectivity into our new day.
Photo Credit: Pixelklex / Shutterstock