© REUTERS/Umit Bektas

1. At its core, systemic competition with China consists of rival visions of human nature. The European ideal of the autonomous individual, responsible for its own actions, has come under severe threat by digitally-driven behavioral conditioning. However, from a Chinese point of view, the Enlightenment view of human nature is a transient historical phenomenon, a luxury that only a few wealthy societies can afford.

In our epoch of interactive digital media, gamification and artificial intelligence, the autonomous individual—capable of thinking, deciding and acting for itself—has been increasingly called into question. That individualistic figure had underpinned the Enlightenment’s optimistic vision of human nature. But today, vast quantities of data amassed by digital platforms and the gaming and advertising industries have proven the extent to which human desires and preferences can in fact be exploited and manipulated, as well as analyzed for their supra-individual properties. It is also increasingly clear how easily most “netizens” can be influenced by the targeted manipulation of information and human emotions. Networks of influencers and followers—in other words, digital thought leaders and acolytes—are by now a powerful vehicle to steer collective mimetic behavior.

To use a rather pointed metaphor: the ecosystems of interactive apps are a playground for modern herd behavior among humans. This herd behavior can be observed in the imitation of prefabricated lifestyles, fashions and looks, but it also extends to individual positions taken on divisive, in-or-out questions, and to forms of collective online denunciation, aggression, and hatred. What confronts us in social media, the gaming industry, and digital advertising are not images of self-aware, mature individuals. We are witnesses to a pandemic of herd behavior.

In a vast number of experimental studies, neuroscience and behavioral economics have demonstrated that herd and mimetic behavior remains a determining element of human nature, notwithstanding the great efforts made by state and society toward individual education, freedom of decision, and self-responsibility. The reestablishment of authoritarian and totalitarian modes of social order has its basis in collective needs for security and order and in herd behavior that is prompted by sheer terror or mass fear. In this way, the digital proliferation of herd behavior benefits the social engineering of the Chinese Communist Party, which it is now deploying with ever greater confidence and championing elsewhere in the world.

2. The CCP’s design for social order is based on a vision of human nature compatible with both Leninist principles and the digital age. Human beings are understood as herd creatures requiring guidance and steering. Under the guidance of a wise, strong Great Herdsman, the herd can be formed into a placid, productive, and technologically innovative society, but only as long as it is protected by vigilant herding dogs and has its agility maintained by loyal junior herdsmen.

China’s communist revolutionaries and the founders of the People’s Republic always rejected Western individualism as a bourgeois smokescreen; they thought close control of the “masses” by a political avantgarde to be essential. However, Mao Zedong and Deng Xiaoping never had access to the digital technologies of power that Xi Jinping relies on today. China’s model for digital civilization consists of an agile hierarchical order, deploying targeted, seamless digital control technologies in order to steer a conflict-prone mass society down well-defined political paths. As it currently stands, this Chinese model offers a worryingly productive combination of political and commercial conditioning with economic and technological agility. This sort of social order appears particularly compatible with the possibilities inherent to our emerging digital civilization. It can potentially serve as an example for other governments and societies also in search of political stability and economic prosperity.

3. Platform corporations like Alibaba or Tencent and innovation champions like Huawei function as authorized “junior herdsmen.” Under close observation by central political authorities, they must pass on unlimited quantities of data to state bodies for evaluation and control.

The metaphor of herd behavior also applies to other specificities of the Chinese context. The system of herd control, which China quite deliberately brought into being, is led by wise herdsmen: the party leader and the central governing ranks of the CCP. For controlling the herd, the central authorities rely on a division of labor between “junior herdsmen” and “herding dogs.” This is necessary since herd society requires carefully-supported pioneers and role models, who for their part remain agile and innovative. Individual pioneering acts of innovation, in particular in the economy, technology, and science, are publicized in so far as they are socially and politically useful within officially-established parameters. In this context, popular online idols created by the system serve to popularize overarching regime goals, including unquestioned nationalism and unconstrained enthusiasm for technology.

From the perspective central authorities, China’s pioneers should on no account be independent, freewheeling thinkers. Instead, they should desire to emerge from the herd, ascending to high-ranking positions in the official order, while never calling political hierarchies into question. Status and rank within the herd are acceptable values; individual power or even political freedom are not. Jack Ma, founder of Alibaba and a model member of the Communist Party, exemplifies this behavioral pattern: a loyal idol for the digital age.

All of this takes place under the watchful eyes of powerful herding dogs (cyber-administrations at every governmental level, digitally-upgraded state security bodies, state-run cloud operating companies, etc.). When in doubt, the authorities will bite quickly and hard to enforce subordination and discipline. The overall guardians are under instruction to allow considerable freedom of action to junior herdsmen who function as pioneers. The search for new grazing grounds (markets) and new nutritional sources (raw materials, products, business models) is of paramount importance to the highest-ranking herders to make the herd system prosper and expand.

4. The Chinese state, well-equipped with digital technologies, is becoming an agile conditioning apparatus impacting both society and the economy.

The Communist Party has powerful instruments at its disposal: AI-based facial, voice and gait recognition and analytic cameras in public spaces, as well as the ubiquitous smartphone, providing a constant supply of data on individual mobility, communications, and transactions. In part, this can be seen as a matter of surveillance and of early recognition of actual or presumed aberrant behavior. This entails an ongoing analysis of the unending data streams produced by all human interactions. On this basis, individual and collective behavior is regulated. Incentives and sanctions are used to shift overall behavior in the desired direction. All individuals and all legal entities, including official bodies, generate permanently updated data sets, which can in turn be used to evaluate and regulate their conduct. Political, social, and economic trustworthiness is no longer generated through human interaction or formal laws and contracts. Rather it emerges from the data streams and evaluations of trustworthiness apps—in other words through state-organized “social scoring.”

“Big Brother” or “Sauron’s Eye” are inadequate metaphors for this system of control. China’s rulers are aiming to have people internalize the system to such a degree that they no longer even perceive how much the Party controls them. Thereafter, inner resistance will no longer exist. External regulation and control at a distance are meant to become a self-evident and accepted part of life. So if the planners in Beijing have their way, a self-operating surveillance system may come into being, eventually meaning almost no need for police. Breaches of the rules will be socially proscribed, with no active participation on the part of state authorities. Each and every person in society must conform and practice self-discipline, otherwise they lose their trustworthiness ranking and are automatically forced out of society. Used in China’s unconstrained way, digital technology allows for constantly intensifying and granular behavioral control not imposed by a superior authority, but rather anchored and executed within the social fabric.

5. AI is a perfect instrument in the hands of a communist party.

Artificial Intelligence offers the possibility of centralized analysis of huge quantities of data, allowing the recognition of patterns and deviations, and systems of centralized regulation, all without active human participation. The particular strength of AI consists of producing optimized regulatory results for the largest possible population—in other words, it can overcome the chaotic, self-seeking individual pursuit of happiness. These ambitions have been among the central goals of communist revolutionaries since the early 20th century. This is why the IT investor Peter Thiel has accurately characterized AI as “communist.”

But after an initial phase of testing and optimization, AI will also be able to function without the deficient human organization of a communist party. For this reason, China can arrange to have its model of social order installed in other countries and societies without having first to develop a local party-state machine. In its enhanced, 21st-century form, AI will reconstitute the communist agenda of a collectivized, conflict-free society, a vision that was an utter failure during the 20th century. Moreover, it will do so much more effectively than any authoritarian party-state, past or present. Thanks to AI, the original vision of early 20th-century communist revolutionaries will at last come within reach: centralized planning, control, and programming of economy and society, aimed at overcoming all social defects and conflicts, both contemporary and historical.

6. The Chinese panopticon currently emerging arose as the realization of a Leninist dream: at any time, anyone can be observed by the central authorities. This will make possible the formation of a “New Human Being.”
Jeremy Bentham first described the idea of the “panopticon,” a method for enabling the perfect surveillance of large numbers of people by a single overseer. In the industrial age, as a result of cumbersome, large organizations, mass deviance from the norm, and many observational blind spots, this vision could never be consistently realized. However, the fusion of new digital technologies with the more traditional control instruments of Communist Party rule is preparing the way for a substantially more effective Digital Leninism that is currently taking shape in China. In the Chinese system of surveillance and conditioning, the individual cannot know with certainty whether he is actually under observation. But he knows that he is visible at all times, always observable: he thus will have to adapt his behavior in a kind of preemptive obedience.

Behind the ambitions of the Chinese social credit system, we can discern the idea of the “New Human Being”: the dream of a civilization in which human behavior is changed to such an extent that it becomes a lasting collective form, entirely compatible with the broader aims of those in charge. The new trustworthiness system—unlike traditional power instruments like mass mobilization and state terror —is being sold to the Chinese people as technological progress, purported to make the individual’s life easier, safer, and more transparent. In fact, the new developments have been hailed by many, who wish to use, for example, a scoring app to find out whether the person they are meeting or the delivery company they are contracting is worthy of trust. In this assessment, however, one’s own judgment plays no further role. The grounds for evaluation now consist of ostensibly objective data profiles.

7. If we do not oppose the penetration of digital manipulation of preferences and behaviors more effectively than hitherto, time will bring the essence of Chinese developments to our own society: a politically and commercially fabricated and regulated herd organization.

With the pandemic of herd behavior in the digital age, the Chinese government’s vision of human nature has concrete potential to gain ground in many other societies. Systemic competition between China and the West—which first played out mainly in economic and technological realms—is now escalating in politics and ideology. That particular conflict focuses on fundamentally differing conceptions of human nature.
If, in the medium term, China turns out to be the superior economic and technological system, with China’s economy possibly double the size of the United States by 2050, it will have undoubted global consequences for 21st-century governance. Chinese views of society and economy as a herd order in need of surveillance and regulation may prove victorious around the world. And China will reveal itself to be the political and economic system that can make use of the whole range of possibilities within digital civilization the most effectively.

For these reasons, systemic competition with China turns on questions that are even more fundamental than political institutions, technological competition, and security threats. It is a question of the roles and rights of human beings in a future society. China is offering an alternative form of social order, aiming at omnipresent digital conditioning and granular behavioral control. This stands in radical contradiction to the vision of human nature dominant in liberal democracies and market economies.
However, the Chinese conception of ordering human beings, society, markets, and government is gaining attractiveness in many developing countries and emerging markets. China can promise comprehensive technological solutions for politically unstable, economically less productive and conflict-ridden societies, in particular in order to reorganize the wildly proliferating metropolitan areas, now often housing more than 15 million people. The tried and tested systemic solutions that China offers for managing megacities can address infrastructure, living conditions, mobility, energy, environment, and security. Chinese surveillance techniques for public spaces and for online communication have already been deployed in at least 18 separate countries.

Europeans criminally underestimate the strength of demand in many countries for Chinese “smart city” models, including infrastructure and security technology. We cannot rule out in advance that the Chinese-designed systems of authority may allow for peace, wellbeing, and even environmental sustainability in an increasingly densely populated and conflict-prone planet. But if current Chinese conceptions and practices of social engineering come to proliferate globally, the ideas of human freedom and self-determination will head for their digital demise.

8. What can freedom-loving societies do to counter the global advance of Chinese-style digital social engineering?

If we Europeans want to defend what we have struggled to achieve over centuries—individual freedom and choice—the digital transformation will prove an obstacle in many areas of life. European data protection regulations are an expression of this conflict. China’s social scoring system and surveillance state offer a counter-model. Even in China, it is not at all certain that social and political conformity can be implemented without resistance, as suggested by Chinese planners. The large-scale technical systems on which Digital Leninism is based are highly vulnerable. And the social reactions to omnipresent surveillance systems and behavior control will likely become manifest only when systems extend over all areas of life and offer no escape, not even within the most mundane scenes of daily life.
But it would be irresponsible to rely on a stumble, even a collapse, of the Chinese system. In the past, the CCP has repeatedly startled us with its capacity for learning and adaptation. For this reason, we cannot pin our hopes on a Chinese collapse. Rather, what we need to work hard at is a purposive and consistent separation between European and Chinese models of digital control in spite of possible economic costs and political conflicts in relations with China. The discussion on the role of Huawei, the Chinese telecoms equipment supplier, in the development of European 5G networks is only the beginning. Subsequently, we must address the worldwide communications, transactions, and data analysis that are enabled in particular by Tencent’s app “WeChat” and Alibaba’s e-commerce empire. Alibaba and Tencent are dominant ecosystems, not only within China itself, but also for the million-strong Chinese diaspora who all send enormous quantities of data to state-controlled server systems in China.

The fight, however, must not only be fought against manipulative uses of data coming from abroad. On the domestic front, we need to take on the predominantly US-based platform enterprises and targeted data services that have come to dominate the markets. Europe can only achieve its much invoked “digital sovereignty” if it prevents unacceptable practices of data collection and behavior manipulation by suppliers inside and outside Europe. Perhaps most importantly, as a precondition to even think about “digital sovereignty,” Europe needs to nourish the emergence of digital business models for its large domestic market that offers competitive services without destroying free societies.

The decisive question is whether European people, societies, and political institutions can find the strength to at least defend and preserve the core areas of freedom of decision in the manipulative world of digital platforms. Many concrete practical measures are on offer and are ready for implementation. Parts of European society and official regulators are pushing back hard against behavioral conditioning and herd behavior. Because why would any rational person want to be reduced to the human equivalent of sheep, lemmings, or even insects?

The post The Dangers of Herd Life appeared first on Berlin Policy Journal - Blog.

© picture alliance/AP Photo/Jeff Chiu

There are many other networks in the world… But the internet is a network that magnifies the power and potential of all others.” When Hillary Clinton, then US Secretary of State, tried to explain the importance of the internet for foreign and security policy in January 2010, many experts thought it was a waste of time. Should this really be a priority of US foreign policy? After all, as the general thinking went, there were more important issues, like the earthquake in Haiti or global terrorism.

Over eight years later, after the revelations of Edward Snowden, the digital offensive of the so-called Islamic State and the debate over “fake news,” hardly anyone is dismissing the internet as a distraction from serious foreign policy—even if many no longer regard the internet as a great blessing. (Clinton herself discovered the Internet’s dark side in the most painful of ways during the 2016 presidential election.) Foreign policy without the internet is no longer conceivable. This shift in opinion should be a warning to all those who today dismiss another technology as irrelevant for foreign policy: Artificial intelligence (AI). Just as the internet has pervaded politics, the economy, and society over the last ten years, AI will in the next ten years appear everywhere and disrupt everything. Any country that tries to ignore this development will lose relevance.

Looking around the Berlin foreign-policy world and beyond, one quickly gets the impression that for many people, AI is still uncharted territory. People either think it’s a new buzzword from California —“Yesterday it was Big Data, right?” Or they spread the word that the world’s going under. Soon, they say, AI will create an all-powerful super intelligence that will try to subjugate humanity, like in the action movie “Terminator.”

Better than Humans

Misperceptions like this are based on an incomplete understanding of what AI actually is. It is often confused with full (strong) Artificial intelligence which is seen as superior to human thinking. But that will remain Science Fiction for the foreseeable future or even forever. The AI of today is better understood as a “collective intelligence”: it is often about the digital extraction of data created by humans. Machine learning, a category into which nearly all important AI technologies fall, usually has two stages. First, neural networks—statistical systems inspired by the human brain—are fed vast amounts of data (for example cat pictures), so that they learn to recognize patterns (what cats look like). Then in the second stage, new data is presented, to which the networks apply what they have learned. Simply put: unlike other software, AI code isn’t written by programmers, but by data.

Thanks to the huge computational power of cloud computing firms like Amazon and Microsoft, AI services can often already recognize objects and language better than humans can. The best facial-recognition software already has 99 percent success rate at identifying faces, though only in laboratory conditions. Language recognition services achieve results that are nearly as good. Other programs can comfortably read scrawled handwriting, once they’ve digested about a hundred pages of it.

Companies can apply basic services like this to complicated tasks, as online-giant Google showed at a conference in California in May. Its latest AI-design, Google Duplex, was able to book a haircut appointment on the phone without the other person realizing that he was talking to a machine. For this demonstration, Google had to combine at least three AI services: language recognition, sentence understanding, and word formation.

AI as a Growth Factor

Big technology firms dominate the AI industry. They possess the most and the best data, programmers, and computer systems. But recently other businesses have begun to make use of AI, too. For example, the clothing retailer H&M uses it to detect fashion trends. Unilever, a producer of household and consumer goods, utilizes AI to evaluate job applicants. The energy company Repsol wants to use AI to make its refineries more efficient, while Siemens is using it to optimize the operation of its gas turbines.

It’s hard to predict how much growth AI will create. But the figures will not be small. Accountants from PricewaterhouseCoopers estimate that by 2030, AI will increase world economic output by 16 trillion dollars—more than China and India combined generate today. That figure is about five times the German GDP. According to Ajay Agrawal, Professor at the University of Toronto and author of the new book “Prediction Machines,” the most important economic effect of AI is that it will sharply reduce the cost of prediction, thus making businesses more productive. Just as electricity made light much cheaper—prices soon became 400 times cheaper than at the start of the 19th century—AI will make it much easier to look into the future.

“AI is like electrical power,” experts say. One day it will be used everywhere. And it’s only a matter of time until that is also the case in foreign and security policy—and in the German foreign ministry. But in what form? In a study released in early 2018, the Berlin think tank Stiftung Neue Verantwortung (SNV) identified three focal points: autonomous weapons, economic effects, and consequences for democracy and society.

Autonomous weapons, which take decisions on their own with the help of AI, are perhaps the most threatening consequence of technological development. They can take various different forms, from automated hacker attacks to self-controlling drone swarms. They raise a number of difficult questions, not least how much control of these systems humans can and should have. Arms experts in the US fear ethical asymmetry above all: countries like China could complete dispense with the “human in the loop,” while Western states potentially refuse to cross this red line.

After long internal discussions, Google decided in May to end its participation in Project Maven, a Pentagon program to develop software that can distinguish between people and things in images made by drones. For Chinese firms like Alibaba and Tencent, there are few concerns about such programs, because these companies are already involved in the “civil-military fusion,” as the government in Beijing calls its close cooperation with tech firms.

AI‘s Impact on Policy

As far as the economic consequences of AI are concerned, the effects on foreign policy are hard to evaluate. The technology could help some countries skip over entire levels of development. China wants to become the world leader in the industry and plans to build an AI-economy worth nearly 60 billion a year by 2030. Other countries will lose out, Germany probably among them. The country is considered a straggler. There are fears everywhere that many jobs will be lost, though such fears are probably exaggerated. The McKinsey Global Institute estimates that only five percent of all professions can be automated away using currently known technology (though machines could do part of the work for more than half of all activities).

The question of economic concentration comes up more and more: Amazon, Google, Facebook, and Apple are already dominant worldwide, and AI could make them even more powerful. Denmark’s decision to send an ambassador to Silicon Valley was dismissed in many capitals as a publicity stunt, but it could prove to be forward-thinking.

AI’s impact on democracy and society will likely present major challenges for foreign policy. The internet has already shown that human rights and technology don’t always fit together. While Hillary Clinton in 2010 praised “the network of all networks,” the NSA used it to wiretap masses of people worldwide, as the revelations of former NSA employee Edward Snowden showed a few years later.

AI makes these “Snowden contradictions,” as the authors of the SNV study call them, even more clear. The technology isn’t just the perfect surveillance tool: video cameras equipped with special chips already follow people automatically. AI can also be used for mass manipulation that goes well beyond the most recent disinformation campaigns. American researchers recently discovered that the Chinese government is the source of nearly 450 million online comments per year whose main goal is to distract. Most of them are still written by humans, but in the future more and more artificially intelligent bots could be used.

As well as such fundamental problems, practical issues arise. Data is the most important raw material for AI. China possesses the world’s deepest data pool, especially when it comes to consumers. The country’s 772 million internet users are open to new things and ideas: many of them don’t carry cash in their wallets and only pay with their smartphones. Other countries, particularly Germany, are much poorer in data for cultural and legal reasons. In the future, data, like other resources, may be managed on a national level. And data protectionism is an ever-growing problem for global firms, as the Financial Times recently reported. The number of laws that forbid firms from exporting data has nearly tripled in the last ten years.

Finally, there’s the question of how the German foreign ministry and related institutions will themselves use AI. A study published in June by the British think tank Chatham House lays out three possible applications for governments: AI could create models of international negotiations, thus simplifying them; it could predict geopolitically important events; and it could help assess compliance with international arms control treaties. At least in the last two cases, we are no longer talking about the future. Record Future, a Swedish-American firm, is already using machine learning for early recognition of hacker attacks and other threats. And software from Palantir, a Silicon Valley firm partially funded by the Pentagon, helps inspectors from the Vienna-based International Atomic Energy Agency (IAEA) in their work in Iraq.

Germany the Laggard

How can politics react to all these challenges? A good AI foreign policy begins with a good AI domestic policy. A number of countries have made this key technology a national priority and published detailed strategy plans. Among them are the US and China, but also France, South Korea, and even smaller countries like Finland. By contrast, the government in Germany—a country that struggles with digitalization in general—has only now begun to seriously engage with the issue.

In terms of the development and use of AI, Germany is average at best. According to a response to a parliamentary inquiry, the federal government spends about 27 million Euros a year promoting AI research, which appears quite low compared to other developed nations, though there are no exact figures for comparison. And with few exceptions, German companies are not at the front of the pack: in early 2018, the Expert Committee on Research and Innovation came to the conclusion that other countries are “much more dynamic” in many areas of AI.

Not being a pioneer means that Germany can learn from other countries’ experiences, the SNV argued in a separate paper published in early June (“Cornerstones of a national strategy for AI”). The federal government, the study said, should have greater ambitions than just better promotion of individual AI technologies. Rather, it “should focus on building and supporting a strong, internationally competitive AI ecosystem.”

In short: we need a stool with lots of legs. Promoting research is certainly one of them, but probably not the most important. More important is a strong base. AI competence can’t only be taught in computer science; it also has to be included in other courses of study. Sufficient computational power and venture capital have to be more easily available. Instead of concentrating on the quantity of data, as China and the US do, Germany should prioritize the quality of data, since smaller quantities of relevant, well standardized data often achieve better results. If the mixture is right, an ecosystem like this will create competitive AI services—faster than would government research programs.

For foreign policy, guidelines for action still need to be written. But some basic principles are already clear. The most important is that going it won’t work. Germany is too small to keep up with the competition on its own. Germany strategy has to be integrated into a European strategy. The most obvious partner is France, which is already much farther ahead in terms of developing and using AI.

Competition for Top Talent

Germany also has to figure out which sort of AI it wants to stand for. There’s a lot of space between China’s state-capitalism approach and the American data monopolies, and that’s where Germany can make its name. It’s not just about the ethics of using AI, but about a new, smart operating system for the data economy. How can markets be organized around this unusual resource? How can personal data be anonymized? Should people be paid for the data they create?

The answers will also have consequences for the labor supply. AI is about computational power and, of course, data, but without a critical mass of data experts, Germany will struggle to keep up with the rest of the world. Attracting and keeping these people is not just a question of salary, although that area can’t be neglected. Even OpenAI, a non-profit organization in Silicon Valley, pays its top scientists almost two million dollars a year. It’s more important that Germany is considered an attractive AI location. There are implications for foreign ministries, too: if you don’t attract or train employees with AI skills, you will become less important.

The foreign-policy impact of the internet, described early on by Hillary Clinton, took a long-time to become clear. But then the internet showed its global political impact with full force. It will probably be the same with artificial intelligence. Foreign-policy specialists should be prepared, if they don’t want to play catchup.

The post Everything is AI appeared first on Berlin Policy Journal - Blog.