21 Lessons for the 21st Century

To the global elite in the 20th century, the liberal story, which enshrines the value and power of liberty, was the best manual over the communist and fascist stories for navigating political, social, and economic challenges. However, the global financial crisis of 2008 has created disillusionment with the liberal story. The rise of Donald Trump in the United States and the Brexit vote by Britain signify this disillusionment has reached the core liberal countries.

Part of this disillusionment stems from liberalism being unable to find solutions to challenges caused by technological disruptions, such as the internet. Liberalism is even more ill-equipped to deal with the twin revolutions in infotech and biotech. For ordinary people, they believe they are becoming increasingly irrelevant because of these disruptions and are losing their economic worth. The breakdown of liberalism has left a vacuum that other creeds are trying to fill. Many of these include “nostalgic fantasies about some local golden past” (14-15).

Harari emphasizes that if liberalism or some other creed wants to continue to play a role in shaping the world, then it will need to make sense of bioengineering and artificial intelligence. It will also need to incorporate these technologies into a new story that unites the global community. Harari offers the following advice for taking the first step towards achieving this goal: Tone down the doomsday prophecies, stop panicking, and work to clarify some of the bewildering possibilities that humanity might soon face with the merging of biotechnology and artificial intelligence.

Machine learning and robotics will change nearly every industry, yet we do not fully understand the nature of this change or how soon it will happen. There are fears that automation will result in massive unemployment. Right now, many manual jobs are being automated, but new service jobs are emerging that require human cognitive skills, such as understanding human emotion, learning, communicating, and analyzing. The concern with these cognitive skills is that artificial intelligence will eventually understand the biochemical mechanisms that underpin human choices, emotions, and desires. Therefore, machines will soon perform these jobs.

Because humans are individuals, it is difficult to make sure that all humans are up-to-date and connected. This is not the case for artificial intelligence, which possesses two important nonhuman abilities: updatability and connectivity. These nonhuman abilities have potential advantages, which might result in computers replacing all humans in some fields. This would even occur when individual humans might still do a better job than machines. For example, AI doctors might provide better and cheaper healthcare, especially to those who currently lack access.

Creativity represents one human trait that currently is difficult for automation. Harari cautions, however, that no job in the long run will be safe from automation. Human emotion defines art. Yet, algorithms are getting closer to understanding and manipulating human emotions. They might even prove more adept at creating artwork that are global hits because they have access to global biometric databases. Through these databases, they understand how biochemical mechanisms produce various human emotions. Most humans are not great artists. Therefore, artificial intelligence will just need to be better than the worst artist. In addition, there are some computer programs that are already showing signs of creativity. AlphaZero, which learned to play chess in four hours, developed chess strategies that no humans have ever used.

The loss of traditional jobs to automation will be offset by the creation of new human jobs. The problem with this scenario is that these new jobs will demand high levels of skill and expertise. There is already a shortage of skilled labor. Because of this, Harari notes, “we might nevertheless witness the rise of a new useless class” (30). Harari again cautions that even these high-skilled jobs will never truly be safe from automation. This means that humans will constantly need to create new jobs and retrain workers. The average human, however, might not have the emotional stamina for life filled with change.

Harari proposes three solutions. The first is that governments work to slow down automation to allow time for society to adjust. Yet, education and psychological changes will also need to accompany this policy for it to be successful. We will need more high-skilled workers who thrive in uncertain and ever-changing environments. It remains unclear, however, whether humans can maintain mental stability in the face of constant change. Thus, governments need to prepare for post-work societies. The first step towards this preparation is acknowledging that all current political, economic, and social models are inadequate for dealing with this global dilemma. Harari argues that we need models that focus on protecting humans over jobs. 

The liberal story believes that humankind is marching towards a global society of democratic politics and free markets. At the heart of this story is the belief that humans are rational agents. Liberalism also assumes that authority stems from human free will, which is expressed through feelings and decisions. This reliance on human feelings, however, might destroy the liberal story. Once artificial intelligence can control and manipulate human feelings, this authority will shift from humans to algorithms.

There is no scientific evidence that free will exists. Biochemical mechanisms, found in all mammals and birds, are responsible for producing feelings to make fast-paced calculations about survival and reproduction. Biometric sensors already convert individuals’ biometric traits into electrical signals that are stored on computers. With enough biometric data and computing power, algorithms could understand an individual’s desires, emotions, and decisions. Algorithms, like humans, will make mistakes, but these algorithms do not need to be perfect. Rather, they need to be better than the average human, which will not be difficult because most humans do not truly know themselves and make mistakes in the most critical decisions of their lives. The temptation to rely on algorithms might increase because they make better decisions than humans about careers, relationships, and possibly even global problems. By gradually giving more and more of the decision-making power to algorithms, humans will inadvertently give them more authority too. 

People argue that algorithms will never make important decisions for humans because many of these decisions involve ethics, which algorithms do not understand. Yet, humans use emotions over ethics to calculate life-and-death decisions. Because computers have not been shaped by natural selection, algorithms might be better at following ethical guidelines in crisis situations than humans (provided we learn to code ethics).

To Harari, humans should fear artificial intelligence because it might give rise to digital dictatorships. Harari notes that one of the strengths of democracy in the late-20th century was that it was better at data processing than dictatorships. The reason for this is that “a democracy diffuses the power to process information and make decisions among many people and institutions, whereas a dictatorship concentrates information and power in one place” (66). Artificial intelligence, however, might destroy liberty. Artificial intelligence stores information in a central place, which could make centralized data-processing systems more efficient than diffused systems. As algorithms store more and more information about humans, authoritarian regimes will truly know everything about their citizens. These regimes might also use algorithms to manipulate how humans feel about them. 

In this chapter, Harari focuses on the role data might play in creating inequality. For thousands of years, the most important human asset was land. Property is a prerequisite for societal inequality. With the domestication of plants and animals, land ownership became increasingly important. A small group of elites slowly monopolized most of the wealth and power, creating rigid social and political hierarchies. The notion of equality is unique to the late modern era, arising out of the industrial revolution and the birth of communism and liberalism. A focus on reducing inequality between classes defines the 20th century. Many believed this decline in inequality would continue in the 21st century. However, inequality is growing around the world, with the 100 wealthiest individuals owning more of the world’s wealth than the poorest 4 billion.

Artificial intelligence might worsen inequality, because it might lead to longer lifespans and upgraded cognitive and physical abilities. However, only the ultra-wealthy will afford these upgrades. The human species itself might then divide into two biological castes: a large underclass of “useless” (75) humans and a small group of superhumans.

To prevent the concentration of wealth and power in the hands of superhumans, Harari argues that we must regulate the ownership of data. One possibility is nationalizing data. While this might curb large corporations’ power, it might also birth digital dictatorships. History teaches us to fear the power of both big corporations and political leaders. Another possibility is private ownership of one’s own data, but this remains unclear. Humans, in contrast to regulating land and machines, do not have much experience regulating data. Data regulation is a much different task than land or machine regulation because “data is everywhere and nowhere at the same time” (80). To Harari, the most important question of our modern era is about how to regulate data ownership.