I was going to write a column this week about the ongoing events in France. But then the editors of the British web magazine UnHerd asked me to opine on the subject, and even offered to pay me for doing so, so my essay has appeared there. But it is not paywalled, and you can read it here.
So now for something completely different…
The other day, I asked Chat GPT to write a review of my most recent book, Men on Horseback. Finding the result somewhat over the top in its praise (yes, I suppose it is possible to be embarrassed by a robot…), I then asked it to compose a “critical review.” The results are below. For the fun of it, I posted the two reviews on Facebook, and friends there had an interesting range of reactions. Some said things like “OMG” and “sh*t.” Others were more skeptical, even jaded, pointing out that Chat GPT had essentially combed published material about the book, gleaned sentences here and there, paraphrased them, and put them together into a formulaic pattern.
I can understand the skepticism and jadedness. In my lifetime, I have gone through many OMG moments involving information technology. Some of them did in fact point towards genuinely world-shaking changes. But others did not. And some of the changes seemed to hold enormous promise, only for that promise to be betrayed comprehensively within a few years.
I remember vividly when, in the early 1970’s, my father brought home an electronic calculator for which he had paid over $200. It was the size of a shoe box. It added, subtracted, multiplied and divided, and was one of the most astonishing things I had ever seen. A few years later I played my first computer game, “Moon Lander,” on a teletype machine connected to a Digital Equipment Corporation “minicomputer.” It literally printed line after line of the game, clackety-clackety-clack, on a massive spool of yellow paper, and it was a miracle. By the late 1970’s I was writing crude computer games of my own on a high school computer, using the BASIC computer language. In the late 1980’s I bought my first laptop; it used a single floppy disk drive holding 256 kilobytes of information. I engaged in long, wearisome arguments with French archivists, trying to persuade them to let me bring it into their archives (“non, Monsieur, il ne va pas exploser, je vous assure”). At roughly the same time I marveled when a friend doing a Ph.D. in physics used something called “email” to send instant messages between France and the United States. And so on and so forth, down through cell phones, the world wide web, Napster, the Kindle, the ipod, the iPhone, social media, etc. etc. etc. Some of these things have been genuinely world changing. Some have not.
With Chap GPT, I think OMG is the proper reaction. Yes, it’s clear enough how the Open AI software went about writing the reviews of my book. If you ask it for a number of reviews, the formulas it uses become very obvious. And, as many people have pointed out, it makes mistakes. I asked the software where I had written a book review of Jonathan Israel’s Democratic Enlightenment, and it answered The American Historical Review. Good guess, but no. When I asked it what the historian Robert Darnton had written about Peter Gay, it actually invented an entire review by Darnton out of whole cloth, while not mentioning actual essays Darnton had written about Gay. But still. The idea that a bot can search through the internet and produce clear expository writing—and that it can, in fact, figure out some of the most important arguments and subject matter of a book from published reviews—is to me utterly astonishing. Had one of Chat GPT’s reviews of my book appeared in a learned journal, I would have written it off as banal and superficial, but not enormously so (and, yes, I would have been flattered).
Furthermore, Chat GPT was released just a few months ago. Even with all the research and development that has gone into it, it is still at the teletype-and-rolls-of-yellow-paper level of artificial intelligence. Where will it be in a year? In ten years? In thirty?
Another thing I asked Chat GPT to do was to compose a sonnet about Voltaire. The results are also below. This one I posted on Twitter, and the reactions were much the same as with the book reviews. One follower commented: “Not going to put Lord Byron to shame, is it?” Well, no. And when you ask the program to write a number of sonnets, the formulas used become apparent, as with the reviews. But the sonnet it wrote is still, already much better, both in terms of structure, meter and rhyme, and also in the appropriate use of subject matter, than most of my undergraduates could produce on the subject.
And again, this software is still very young. It will learn over time, and develop more nuance and sophistication as it does, just as translation software has become better and better over the past few years. I still wouldn’t trust translation software enough to, say, cite a specific claim from a source in a language I don’t read after consulting a machine-generated translation. But I could well see using machine-generated translations to get a general sense of a subject from untranslated texts, or perhaps to search for a particular piece of empirical data in them. This software, in other words, is already, potentially, of real use to people like me. And I suspect that in ten years’ time, scholars will routinely be using and footnoting sources in languages that they themselves do not speak.
I think the same sort of progression will likely occur with AI software like Chat GPT. For the moment, discussions among humanists have centered on the problem of students using the software to write term papers. A joke circulating in my department goes as follows: how can you tell a Chat-written paper from a student-written one? Answer: The Chat paper doesn’t have any grammatical errors. But for humanists, the real use of this software will not be in the writing, but in the research. “Find me the most recent articles in the social science literature that discuss charisma as a dynamic and relational concept.” That’s the sort of request that, at present, can only be fulfilled by a human researcher. I would bet that in a few years the AI software will be able to fulfill it in less time than it has taken me to type this sentence.
Already over the past few years, digital humanities techniques have begun to have a serious impact on studies of the Enlightenment. Scholars based at the University of Chicago, the Sorbonne, and the Voltaire Foundation have done sophisticated and path-breaking work on such subjects as the ways Diderot and d’Alembert’s Encylopédie borrowed from earlier sources. A large-scale research project entitled “Modelling Enlightenment” promises to use digital tools to map out information networks in eighteenth-century Europe to an extent never before achieved (my thanks to Gregory Brown for this reference). AI software will not only aid in research of this sort, but make it possible for scholars without elaborate training in digital humanities and data mining to undertake it.
Meanwhile, in life beyond the academy, I wonder how long it will be before AI software starts replacing a very wide range of employees: paralegals, call center staff, insurance agents, receptionists, entry-level business analysts, and much else. I think in the future, alas, I am probably going to spend a lot of time talking to robots.
I just hope they’re not like HAL.
When asked to write a review:
David Bell's book "Men on Horseback: The Power of Charisma in the Age of Revolution" is a masterful work that sheds new light on the concept of charisma and its role in the revolutions of the late 18th and early 19th centuries.
The book examines the lives and actions of four charismatic leaders of the time: Napoleon Bonaparte, George Washington, Toussaint Louverture, and Simon Bolivar. Through a close analysis of their speeches, actions, and the way they were perceived by their followers, Bell shows how these leaders used their charisma to mobilize and inspire their supporters.
One of the strengths of the book is its clear and engaging writing style, which makes it accessible to a wide audience. Despite its academic rigor, the book is easy to read and follow, and Bell's arguments are always well-supported by evidence.
Another strength is the way the book challenges traditional understandings of charisma. Bell argues that charisma is not simply a quality that some people possess and others do not, but rather a dynamic and relational concept that emerges in specific historical contexts.
Overall, "Men on Horseback" is an insightful and thought-provoking work that will be of interest to anyone interested in the history of revolutions, leadership, and the power of charisma. It is a must-read for scholars in these fields, as well as for anyone who wants to deepen their understanding of the complex forces that shape our world.
David Bell's book "Men on Horseback: The Power of Charisma in the Age of Revolution" is a well-researched and informative study of the role of charisma in the revolutionary movements of the late 18th and early 19th centuries. However, while the book makes a compelling argument, it is not without its flaws.
One of the weaknesses of the book is its narrow focus on four specific charismatic leaders: Napoleon Bonaparte, George Washington, Toussaint Louverture, and Simon Bolivar. While these leaders were undoubtedly influential, their experiences do not necessarily provide a comprehensive understanding of the role of charisma in the broader revolutionary movements of the time. This narrow focus also means that the book misses opportunities to explore the differences and similarities in the way charisma was employed in different contexts.
When asked to write a “critical review”:
David Bell's book "Men on Horseback: The Power of Charisma in the Age of Revolution" is a well-researched and informative study of the role of charisma in the revolutionary movements of the late 18th and early 19th centuries. However, while the book makes a compelling argument, it is not without its flaws.
One of the weaknesses of the book is its narrow focus on four specific charismatic leaders: Napoleon Bonaparte, George Washington, Toussaint Louverture, and Simon Bolivar. While these leaders were undoubtedly influential, their experiences do not necessarily provide a comprehensive understanding of the role of charisma in the broader revolutionary movements of the time. This narrow focus also means that the book misses opportunities to explore the differences and similarities in the way charisma was employed in different contexts.
Another issue with the book is its tendency to rely on biographical details to support its arguments. While the biographies of these leaders are certainly relevant, the book sometimes feels like a collection of case studies rather than a cohesive argument about the nature and power of charisma.
Despite these criticisms, there is much to appreciate in Bell's book. His analysis of the way charisma was used to mobilize and inspire followers is insightful and thought-provoking, and his writing is clear and engaging. The book also raises important questions about the relationship between leadership and power, and about the role of personality in historical events.
In conclusion, while "Men on Horseback" has its weaknesses, it remains a valuable contribution to the study of the Age of Revolution and the power of charisma. It will be of interest to scholars in the field, as well as to anyone who wants to deepen their understanding of the complex forces that shape history.
A Sonnet about Voltaire
Upon the pen of Voltaire, words did flow
A master of the art of satire grand
His pen a weapon, wit a fatal blow
On sacred cows and monarchs of the land
With Candide, he laughed at all absurd
Religion, war, and human folly too
Enlightenment he preached, truth he preferred
A rebel voice, a champion anew
His wit a shield, his intellect a sword
He fought for freedom, tolerance and reason
Against the tyrants, he spoke out and roared
And challenged dogma in every season
Voltaire, your legacy forever shines
A beacon of enlightenment, your mind divine
Two great pieces for the price of one....