She’s a key figure behind the rise of artificial intelligence today, but not all computer scientists thought Fei-Fei Li was on the right track when she came up with the idea of ImageNet, a massive visual database to build on. It took years.
She’s a key figure behind the rise of artificial intelligence today, but not all computer scientists thought Fei-Fei Li was on the right track when she came up with the idea of ImageNet, a massive visual database to build on. It took years.
Lee, now the founding director of Stanford University’s Institute for Human-Centered Artificial Intelligence, is out with a new memoir that chronicles his pioneering work in validating the dataset that gave rise to the computer vision branch of AI. Did you speed up?
The book, “The World’s IC,” also charts his early years as he suddenly moves from China to New Jersey and follows him through academia, Silicon Valley and the halls of Congress to the rise of AI technology. Commercialization drew public attention and response. He spoke with The Associated Press about the book and the current AI moment. The interview has been edited for length and clarity.
Q: Your book explains how you envisioned ImageNet as more than just a huge data set. Can you explain?
A: ImageNet is a really cool story of identifying the North Star of an AI problem and then finding a way to get there. North Star for me was really about rethinking how we can solve the problem of visual intelligence. One of the most fundamental problems in visual intelligence is understanding or seeing objects because the world is made up of objects. Human vision is based on our understanding of objects. And there are many, many, many of them. ImageNet is really an attempt to define and provide a way to solve the problem of object recognition, which is a big data path.
Question: If I could travel back 15 years to when you were working hard at ImageNet and tell you about DALL-E, Stable Diffusion, Google Gemini and ChatGPT, what would surprise you the most? ?
A: What doesn’t surprise me is that everything you mention — DALL-E, ChatGPT, Gemini — is based on big data. They are pre-trained on a large amount of data. That’s exactly what I was hoping for. What surprised me is that we got creative AI faster than most of us thought. For humans, race is actually not that simple. Most of us are not natural artists. The easiest generation for humans is words because speaking is creative, but drawing and painting are not creative for normal humans. We need the Van Goghs of the world.
Q: What do you think most people want from intelligent machines and does it align with what scientists and tech companies are building?
A: I think basically people want respect and a good life. This is almost a fundamental principle of our country. Machines and tech must be compatible with universal human values – dignity and a better life, including freedom and all that. Sometimes when we talk about tech or sometimes when we create tech, whether it’s intentional or unintentional, we don’t talk about it enough. When I say ‘we’ it includes technologists, it includes businesses, but it also includes journalists. This is our collective responsibility.
Q: What are the biggest misconceptions about AI?
A: One of the biggest misconceptions about AI in journalism is when journalists use the subject AI and a verb and put humans in the object. Human agency is very, very important. We build technology, we deploy technology, and we govern technology. The media and public discourse, but heavily influenced by the media, is talking about AI without due respect for human agency. We have a lot of articles, a lot of discussions, starting with ‘AI blah, blah, blah’. AI does blah blah blah AI provides blah blah blah. AI destroys blah, blah, blah.’ And I think we need to recognize that.
Q: You studied neuroscience before getting into computer vision, how different or similar are AI processes to human intelligence?
A: Since I’ve scratched the surface of neuroscience, I have an even greater respect for how different they are. We don’t really know the intricate details of how our brains think. We have some traces of lower-level visual functions such as seeing colors and shapes. But we don’t know how humans write Shakespeare, how we fall in love, how we designed the Golden Gate Bridge. There are just so many complexities in human brain science that it is still a mystery. We don’t know how we do it in less than 30 watts, the energy the brain uses. How come we’re so terrible at math when we’re so good at seeing and navigating and manipulating the physical world? The brain is an infinite source of inspiration for what artificial intelligence should be and do. His neural architecture — (Nobel Prize-winning neurophysiologists) Hubel and Wiesel were actually its discoverers — was the beginning of the artificial neural network movement. We borrowed this architecture, although mathematically it does not fully mimic the workings of the brain. There is a lot of interconnected inspiration. But we also have to respect that there are a lot of unknowns, so it’s hard to answer how similar they are.