It is nearly impossible today to read an article, watch TV, or scroll your social media feeds without a dozen references to artificial intelligence. Many of us use AI daily at our workplaces or are involved in the development of the next era of neural networks, the core of the AI world. We understand how AI can benefit society and engineering companies. Yet for many, AI represents a confusing world filled with potential dangers.
The actors’ and writers’ strikes in Hollywood last year were just two of the recent occurrences highlighting the potential misuse of AI. Although several issues led the Screen Actors Guild and the Writers Guild of America to strike, the issue that rose to the forefront was the use of AI to scan an actor’s image for use in future projects and AI-generated writing as source material. “This is the first step on a long process of negotiating and working through what generative AI means for the creative industry — not just writers but visual artists, actors, you name it,” said David Gunkel, a professor of media studies at Northern Illinois University and author of Person, Thing, Robot, in an AP news story about the strike.
As I began reading The Worlds I See by Fei-Fei Li, Ph.D., I discovered that it is an autobiography told in conjunction with the evolution of AI. It was a refreshing and enjoyable excursion into the history of AI and Dr. Li’s life and career.
Dr. Li was born in China and immigrated to the United States with her parents in 1992 at the age of 15, speaking limited English but with a passion for physics and learning in general. They landed in New York City as the digital revolution in the United States was underway. Dr. Li begins her description of the development of AI with the historic Dartmouth College summer research project in 1956 that coined the term “artificial intelligence.” The idea was proposed in a paper published by the college in 1955 that stated:
We propose that a 2-month, 10-man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.
As Dr. Li learns what it is like to be a student in a U.S. high school and experiences the challenge of learning a new language, she also helps her parents assimilate in this new world and culture. With the help of her math teacher, a future friend and mentor to her and her family, Dr. Li earns a scholarship to Princeton to study math and physics. She also has an interest in the study of computation, once again needing to learn a new language. She develops an interest in how the brain learns new things and processes thoughts in a way that is similar to computers. “It was dawning on me that silicon wouldn’t just help us decode the nature of the mind but help us model it.”
AI was advancing during this same time, with computers learning to read handwritten digits such as ZIP codes and neural networks using backpropagation to train computers to reprogram networks to improve accuracy. But a divide surfaced between the theory and practice of machine learning. A period of time emerged in the history of AI, referred to as “AI winter,” in which the broader concepts of AI were downplayed as researchers studied pattern recognition and decision-making.
As Dr. Li continued her education at California Institute of Technology in electrical engineering and computational neuroscience, she explored the science of human vision to understand how a computer might “see” an image, evaluate what it is seeing, and categorize it as humans do — as people, places, and things. Her research focused on exposing computers to hundreds of images and developing algorithms to learn new categories of objects.
Dr. Li continues to parallel her life with the development of AI, masterfully interweaving text about her colleagues, family, and friends with the continuing history of AI and the numerous academics and scientists involved in its development. The creation of an immense dataset called ImageNet becomes her “North Star,” the goal of a lifetime. It is made possible by advances in the internet, digital cameras, and the ability of computers to operate at faster speeds. The hypothesis was “that the first step toward unlocking true machine intelligence would be immersion in the fullness of the visual world.”
The book was easy to read; at times I was unable to put it down because I was eager to see what came next, not only in Dr. Li’s life, but equally in AI.
I attribute my understanding of the AI topics covered to my own career, ranging from my first year as a computer science major to becoming an engineering faculty member to retirement. As faculty, I listened to students defend theses on neural networks and machine learning, and as a retiree, I discovered that much of the technology I interact with daily has a component of AI in it.
This book clarified much in my quest to better understand AI, its origins, its ability to benefit, and its ability to harm. As Dr. Li states, “The future of AI remains deeply uncertain, and we have as many reasons for optimism as we do for concern.”
References
Fei-Fei Li, Ph.D. (2023). The Worlds I See, Macmillan Publishers.
Dartmouth Summer Research Project on Artificial Intelligence (1956).
David Gunkel (2023). Person, Thing, Robot, MIT Press.
Associated Press (Sept. 5, 2023). “In Hollywood writers’ battle against AI, humans win (for now).”
