Artificial intelligence may seem like a state-of-the-art advancement with roots only beginning to take hold.
Not so, according to Kevin Donnelly, Ph.D., associate professor of history.
In May, Donnelly released his second book, “The Descent of Artificial Intelligence: A Deep History of an Idea 400 Years in the Making.” Itself seven years in the making, the book was published by the University of Pittsburgh Press.
The idea for the book began percolating when Donnelly spoke at a 2017 conference at Penn State’s Center for Humanities and Information, which featured presentations on statistics, data collection and algorithms throughout history, the types of information that make up today’s AI.
“It struck me that while ‘artificial intelligence’ had been around as a term for over 70 years — and many futurists and tech people were starting to predict when AI might reach a human level — no one was really looking at what kind of human AI was supposed to replicate,” he said. “To understand how AI might work, there needed to be a history of how people have thought, or at the least how people had tried to define how we think.”
In the book, Donnelly traces AI’s origins to the popularity of the ideas of René Descartes in Europe in the 17th century. Mechanical philosophy imagined that people are subject to scientific understanding just like other parts of nature, and that the same kind of experiments, quantification and scientific methods applied to nature could be applied to people.
Other big historical events profiled in the book are the Enlightenment, the French Revolution, the Industrial Revolution and the Cold War. In terms of scientific thinking, Isaac Newton, Pierre-Simon Laplace, Charles Darwin and Alan Turing provide the big ideas.
“While the book is not at all a history of these events and people, it shows how major historical events in Europe and America led thinkers — particularly social scientists who study people as if they were machines — to imagine the ‘human level’ in new ways,” he said. “I argue that each new scientific description that emerged from these thinkers simplified people, making the human level something closer to a machine. While many histories discuss the rise of AI, focusing on the increasing complexity of machines, the book looks at the lowering of human complexity. This is where the ‘descent’ part of the title comes in.”
One of society’s most common misconceptions about AI, according to Donnelly, is the perception of it as a kind of effortless technological marvel. In fact, it resulted from enormous amounts of human labor.
Should people be concerned about AI?
“Yes, but probably not for the reasons portrayed in dystopian films, Silicon Valley and in the tech world, where people openly discuss the percentage chance of something like an extinction-level event,” he said. “The major concerns I think are closer to those that people already have, or should have, with the power concentrated in major American technology companies. The fact that most AI today is tied directly to ideals and goals of profit-making companies rather than democratically elected governments is deeply problematic.”
Donnelly has used ChatGPT to test various questions and prompts to detect cheating on student papers and exams. But he doesn’t use AI in his personal life, at least not consciously.
He believes one of his book’s strongest attributes is that it gives people the chance to view artificial intelligence from the perspective of humans. Only a select few can understand AI at the level of pure technology, which can lead to the pessimistic belief that humans are at the mercy of technology and the smart people who create it.
“Fortunately, the book demonstrates that AI today is also the result of the stories we tell about ourselves,” he said. “And the one thing we can be in control of is how we define ourselves.”
“The Descent of Artificial Intelligence: A Deep History of an Idea 400 Years in the Making” is available on Amazon, Barnes and Noble and other online booksellers.