Skip to main content
ResearcherCautiousTier 2

Fei-Fei Li

Co-Director, Stanford Institute for Human-Centered AI (HAI), Stanford University

The researcher who taught AI to see — ImageNet sparked the deep learning revolution, and now she's steering AI toward serving humanity.

Credentials

Sequoia Professor of Computer Science at Stanford, Co-Director of Stanford HAI, Creator of ImageNet, former VP at Google Cloud (AI/ML), Member of the National Academy of Engineering, National Academy of Medicine

Why They Matter

Li created ImageNet, the dataset that proved deep learning actually works — without her, ChatGPT might not exist for another decade. Now she's the leading voice for "human-centered AI," pushing the idea that AI should augment workers, not replace them. For business owners worried about AI taking jobs, her framework offers a practical middle path.

Positions

AI Timeline View

AI is powerful but narrow. True general intelligence is still far off. The immediate priority is building AI that works well with humans, not replacing them.

Safety Stance

Cautious

Key Beliefs

AI must be developed with human needs at the center — technology for technology's sake leads to harm.

Stanford HAI founding principles and her book "The Worlds I See"

Diversity in AI research teams directly improves the quality and fairness of AI systems.

Multiple keynotes and Stanford HAI publications

AI in healthcare is one of the highest-impact applications — it can improve diagnostics and reduce inequality in access to care.

TED Talk, "How we teach computers to understand pictures"

Large-scale, high-quality datasets are more important than clever algorithms. Data is the foundation of AI progress.

ImageNet paper and subsequent talks

Controversial Take

Pushed back against the AI doomer narrative, arguing that excessive fear-mongering could lead to overregulation that concentrates AI power in the hands of a few large companies and governments — hurting exactly the people AI should help.

Track Record

How well have Fei-Fei Li's predictions held up?

Large-scale visual datasets would unlock a new era of computer vision and AI

Made: 2007

ImageNet (launched 2009, competition from 2010) proved this spectacularly. AlexNet's 2012 win on ImageNet launched the deep learning revolution.

Right

AI will transform healthcare by augmenting doctors, not replacing them

Made: 2017

AI diagnostics tools are now deployed in hospitals worldwide, but the augmentation-vs-replacement debate continues.

Partially Right

Key Quotes

If we want machines to think, we need to teach them to see.

TED Talk (2015-03)

There's nothing artificial about AI. It's inspired by people, it's created by people, and — most importantly — it impacts people.

TIME Magazine interview (2023-09)

I started ImageNet with a very simple hypothesis: if you want a better algorithm, you need better data. The algorithm will follow.

Stanford HAI lecture (2019-04)

The question is not whether AI will change the world. The question is who is going to be building it and what values they bring.

World Economic Forum, Davos (2024-01)

Publications

Book

The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI

2023

Paper

ImageNet: A Large-Scale Hierarchical Image Database

2009

Last updated: 2026-03-26

Back to AI Minds Directory