“There is nothing artificial about AI. It is made by humans, intended to behave like humans, and ultimately to impact humans' lives.”
“If we don't get women and people of color at the table, we will perpetuate bias and build a world that is not representative.”
“We need red lines in AI. Some things should simply not be built.”
“AI is a dual-use technology. The same tool that diagnoses cancer can be used for surveillance.”
“I worry that AI is being shaped too narrowly by too few people in too few places.”
Fei-Fei Li is the intellectual architect of human-centered AI, a framework that insists technology development must begin and end with consideration of human impact. Her creation of ImageNet in 2009—the dataset that catalyzed modern computer vision—gives her unique standing as both a pioneering researcher and a thoughtful critic of the field she helped build.
Li’s philosophy rejects the framing of AI as an autonomous force, instead emphasizing that AI systems reflect the values, biases, and blind spots of their creators. This perspective drives her advocacy for diversity in AI teams and her insistence that technologists engage with humanists, ethicists, and social scientists.
Through Stanford HAI, Li has built an institutional home for interdisciplinary AI research that brings together computer scientists, economists, philosophers, and policy experts. This model has influenced how universities worldwide approach AI research and education.
Her public statements consistently emphasize “red lines”—applications of AI that should not be pursued regardless of technical feasibility. This principled stance distinguishes her from the “build it and figure out governance later” approach common in industry.
Li’s memoir “The Worlds I See,” published in 2023, brought her story of immigration, perseverance, and scientific discovery to a mainstream audience, further amplifying her influence on how society understands the human side of artificial intelligence development.