Fei-Fei Li

Fei-Fei Li created ImageNet — a dataset of 14 million hand-labelled images across 22,000 categories, assembled between 2007 and 2009 using Amazon Mechanical Turk and the efforts of thousands of annotators globally. The ImageNet Large Scale Visual Recognition Challenge (ILSVRC), which she ran from 2010, became the benchmark that defined computer vision for a decade. When AlexNet won ILSVRC 2012 and halved the error rate, it was competing against a challenge that Li had built.

Without ImageNet, the deep learning revolution would have had nowhere to land. The availability of a large, labelled dataset enabled training at a scale that earlier researchers had not attempted. Li's contribution is often underestimated because it is infrastructure rather than algorithm — but infrastructure is what made the algorithms work.

Li served as Chief Scientist of AI and Machine Learning at Google Cloud from 2017 to 2018 while on leave from Stanford, making her one of the few academics to operate at that level of the commercial AI industry. She returned to Stanford and co-founded the Stanford Human-Centered AI Institute (HAI) in 2019, focused on AI research that addresses human needs and respects human values.

Li has been a consistent advocate for AI research that centres people — particularly through her AI4ALL initiative, which works to increase diversity in AI through education programmes for underrepresented high school students.