Jill Lepore's book “The Government of Machines”: Critique of Automation

How bad will it be? Jill Lepore is worried. Not about a single political project, a specific paragraph, a turn of events in world events, a person, a ban, a product, a publication, an aberration - she sees liberal democracy itself shaken to its foundations. Because, in their opinion, technology and technology are developing too quickly, because this gives too much influence to a few, and because decisions are becoming increasingly irrelevant. Because all of this reduces the members of a community to mere measurable numbers and thus, in a certain sense, strips them of their personality.

The well-known Harvard historian presents a constructive criticism of digitalization in her book “The Government of Machines: The Rise and Fall of the Artificial State”. She writes about the beginnings of political statistics, why knowledge creates states, about how computers and the Internet came into being, about the great technology debates of the post-war period, about what really happened and what leading figures speculated in various literary forms.

An “Electronic Democracy”?

Elon Musk and his grandfather, Jeff Bezos and Mark Zuckerberg appear, John Stuart Mill, Karl Marx, Ayn Rand, Mary Shelley, Hannah Arendt, Joseph Weizenbaum, Isaac Asimov, Douglas Adams, Yoneji Masuda, Alan Turing, Joseph Licklider, John McCarthy, Norbert Wiener, Newt Gingrich, Stewart Brand, Bill Clinton, Donald Trump and many others. Lepore describes closely, sometimes too closely, how our technological environment was conceived and created - and what, in her view, the highly dangerous consequences this has and will have. Their bond is something like the dialectics of automation. It comes at just the right time, as the daily discussions about the dangers of artificial intelligence show, in which the heads of the most important AI companies are now also acting as warnings.

What Lepore is ultimately afraid of was concisely invented by Isaac Asimov in 1955. The science fiction author, best known for his Foundation cycle and his robot stories, published the short story “Franchise” that year. It takes place much later, in 2008, when the American elections are very different. No longer do millions of people vote on their leaders, but only one person. But not just anyone. A person who is particularly statistically representative - in the story his name is Norman Muller, an inconspicuous citizen from Bloomington in the American state of Indiana. A high-performance computer called Multivac calculates how the entire population would probably vote based on how Muller thinks politically and assesses the economic and social situation. That's enough. This replaces the choice. This “saves” Muller’s fellow citizens from having to go to the polling station themselves and cast their vote.

In the Asimov story, the United States is an “electronic democracy.” Lepore quotes an old man named Matthew who can still remember what it was like when he voted. "I went straight into the voting booth, pressed the lever and cast my vote," he tells his ten-year-old granddaughter Linda, and continues: "There was nothing there. I was just saying: This guy is my husband and I'm voting for him. That's how it should be!" His granddaughter, a child of "electronic democracy", asks him: "How did people know who to vote for? Did Multivac tell them?" Her grandfather responded: “You were just using your own judgment, girl!”

Algorithms replace public discussion

And then, and this is a strength of this science fiction, Matthew explains how this actually happened, why the forecasting mainframe suddenly took the place of mass voting. Not with any bad intentions. Out of an idea of ​​efficiency, also out of pure convenience. "You know, Linda, sometimes they had to count each vote all night long, and people got impatient. So they invented special machines that could compare just a few of the new votes with the votes from the same precinct from previous years. In this way, the machine could calculate what the overall result would be and who would be elected. Do you understand?" She understands. And keep listening to him. "The first computers were much smaller than Multivac. But the machines kept getting bigger and were able to predict the election result after fewer and fewer votes. Then they finally built Multivac, which only requires one voter."

Granddaughter Linda, now in her own world of experience, says: “That’s nice.” Grandpa Matthew replies: "No, that's not nice. I don't want a machine telling me how I would vote just because some joker in Milwaukee says he's against a rate increase. Maybe I'd like to vote crazy just for fun. Or maybe not at all. Maybe..."

What robots want

Like a magnifying glass, this Asimov story makes it clear what Lepore means when she writes about an “artificial state”. Human judgments, interpersonal conversations, public debates, parliamentary votes and legal disputes are replaced by digital infrastructures, algorithms and automation. Computers constantly evaluate an almost infinite number of data points, identify patterns and predict who wants to see, hear and feel what. Statistical prediction displaces political majority decisions, politics as a whole becomes a mechanically measurable optimization problem. Not following a large master plan, but due to individual developments that intertwine at some point, but often only in retrospect do not appear coincidentally at the same time.

In the first part of the volume, Lepore traces the historical origins of all this. It begins with the fact that every form of statehood reflects the knowledge available about its members. Counting, classifying and identifying people is what makes political action in large groups possible. The more detailed the data profiles are, the easier it is to influence things - for better or for worse. She then presents in detail the idea, which was increasingly widespread even before the Second World War, of filling important political positions with engineers and natural scientists, because the question of the best political solution essentially has a technical answer. As is well known, the technocracy movement of the 1930s in North America even resulted in the founding of corresponding parties and programs. A well-known representative was Joshua Haldeman, a chiropractor, pilot and adventurer who once emigrated from Canada to South Africa - and later became Elon Musk's grandfather.

When Lepore writes about the leading figures in Silicon Valley today, she draws parallels to this movement and finds comparable, sometimes identical motives. One of these is to overcome traditional democratic decision-making and public administration, which is perceived as too inefficient, slow or lacking in results, with something better. Reference is the result, not the path to get there. An important instrument is the increasing entrepreneurial organization of political institutions - large private-sector internet platforms create publicity, but often reward less the objective exchange and facts than the manipulative exaggeration, outrage or fiction.

The rise of the artificial state

She problematises how much political power lies in the hands of a few entrepreneurs. And contrast this with how, for example, the Internet was once conceived. How the influential Californian activist Stewart Brand was not the only one who imagined a decentralized, small-scale digital space, far away from Washington, in which all people would be more equal than they actually ever were. And how the USA later, under President Bill Clinton and leading congressman Newt Gingrich, introduced internet legislation that led to the participatory web with its large platforms, which today connects everything with everything.

The second section of her book, entitled “What Robots Want,” makes extensive reference to the science fiction and philosophy that deal with the artificial replication of humans and the coexistence of human-intelligent machines with humans. On the one hand, Lepore uses this to explain what the intellectual superstructure of the inventors and technicians was and is. What inspired, what deterred, which perspective continued to have a scientific impact (such as Descartes' view of animals, life and thought). And how technology, power, control and loss of control were thought together, right up to the speculation about whether there is a risk of extinction, because such warnings have not only been around since large AI models were able to use language competently.

Lepore's recourse to the literary is also due to the fact that, despite all the progress, there are currently no humanoid robots like C3PO or Commander Data, no AI that can sustain itself and act autonomously in the real world. Lepore needs speculation, the fictitious, to substantiate some warnings because the empirical evidence is missing.

Their descriptive conclusion is accordingly devastating: "The artificial state in which very many people have now begun to live has not promoted their security and happiness; it is a prison whose walls are made of mirrors: glowing screens. In its cells there are no windows that open to the world outside, to the sea and the sky, to forests, valleys, islands and cliffs. From within you cannot hear a sound, not the rustling of animals, not the snapping of branches, not the roar of the waves. No vines creep under the door. No spider web stretches in the corners. Your forecast conclusion is nevertheless confident. This artificial state will not last. An “electronic democracy”, as Asimov imagines it in his short story, is not the end of our story, that is certain. The failure was basically already inherent in the rise of the artificial state.

This is exactly how this book should be understood: Lepore has written an informative warning at the right time. One that fortunately doesn't stop at this or that new AI functionality, interface, data center innovation or Trump tirade. It is true that it idealizes the intellectual counterpart to its artificial state, liberal democracy, and essentially ignores the beneficial consequences of technical progress, but it is not relevant for this purpose. More and more observers and citizens share their feeling that something is going wrong, that humanity has reached a technological tipping point. In some ways, Lepore is perhaps like the hero Hari Seldon from Asimov's Foundation series - she cannot avert a coming catastrophe, but shows a way to overcome it more quickly.