Fei-Fei Li is a computer scientist at Stanford University whose research focuses on artificial intelligence and computer vision. She co-directed the ImageNet project, helped advance modern machine-learning research and serves as founding co-director of the Stanford Institute for Human-Centered AI.
[This interview was edited for length and clarity.]
How would you describe the current state of American science?
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
America remains the most vibrant country when it comes to scientific innovation. The ecosystem is still the healthiest in the world in the sense that the amount of private investment, as well as public investment, in basic science research and industry science research is really high.
Having said that, I think America is changing. Our scientific federal investment hasn’t increased for many, many years, and new technologies such as AI are extremely resource-heavy, including compute resources, data resources and talent. All of these three resources, right now, are rapidly being concentrated in private sectors, which, on one hand, continues to move America’s innovation forward, which is a great thing. A lot of good technologies are coming out of our tech labs.
On the other hand, this is not sustainable in the long term. Basic science remains very critical for the health of the ecosystem, not only from a talent-training point of view but also from a knowledge- and curiosity-driven research perspective that lays the foundation for deeper tech transfer and innovation for generations to come. So while I am a believer in America’s science ecosystem, I am concerned that we should pay more attention to the health of this ecosystem and try to seize this moment.
What needs to change in American science?
I already mentioned the resourcing of the ecosystem and public sector. I also think a healthy science ecosystem starts from K–12 education, where STEM education should be critically resourced and revamped. America continues to attract many talents from across the globe, but as the geopolitical landscape evolves, as well as the limitation of immigration, we need our own in-house, in-country talents for the future of our science. These talents right now are in K–12. Our teachers, as well as our students, especially in the public education system, are underresourced, and STEM has been an area where we have not made significant progress in the past decades. America as a country ranks a lot lower than most developed countries in terms of math evaluation for K–12 students. These are alarming warning signs.
In addition, I think, in the 21st century, we see a new type of civilizational technology coming of age, which is artificial intelligence. AI is an extremely horizontal tool that would empower cross-disciplinary science—from biology to medicine to chemistry to physics to energy to climate to social sciences, from economics to political science to history to psychology, and many more [disciplines]. We need to encourage and resource this kind of cross-disciplinary science.
I also think America’s higher education system is facing change. It has grown in size a lot, including the bureaucracy across the board. We hear this from every university we visit. And with this new set of tools in AI, how to run our higher education system needs critical rethinking.
What gives you optimism right now?
My optimism is rooted in our system. We have a robust social system rooted in democracy and freedom, and this is very important for scientific quests. Scientists need to have stable environments where they can pursue their curiosity, sometimes for decades to come. We also have an incredible higher education system where curiosity is maximally protected, if resourced well.
What’s your best advice for an early-career scientist?
Scientists are people who are curious and want to pursue knowledge and pursue truth, and this still remains the most timeless and priceless value of a scientific career. My advice to an early-career scientist is to remain true to your passion and curiosity and that sense of mission to discover knowledge and uncover innovation. It’s also very important to be using the tools of the moment, of the century. There are a lot of new tools to uplevel our own scientific research and scientific quest. No matter which discipline an early-career scientist is in, from medicine to biology to physics, knowing how to use tools such as AI, coding agents and digital research assistants is just fundamentally important.
How has your field changed in the last few years?
My field is AI, which is an extremely vibrant field. Not only has the field itself changed in the past few years, it is the changing force of many other fields. Within our field, we have grown incredibly in size compared with even five years ago. The number of students and researchers working in AI has just continued to multiply, and they are also becoming younger and coming from all walks of life, globally. If you look at conferences such as CVPR [the IEEE/Computer Vision Foundation Conference on Computer Vision and Pattern Recognition] or NeurIPS [the Conference on Neural Information Processing Systems], attendance has gone from low-digit thousands all the way to low-digit tens of thousands. This is a 10-fold increase over the past five years. As the size of the field grows, the output of the field has grown, too. The pace of change in terms of technology is just breathtaking. It is also going through a phase transition. Some of the technologies are so mature that suddenly software engineering is completely disrupted by AI, and we’re going to see this in other parts of our research.
AI has become an enabling technology across other disciplines, empowering technology for other fields, which makes AI a lot more interdisciplinary. In 2026 we are seeing collaborations across every field on a university campus with AI and the usage of these tools. This is transforming AI from a niche computer science discipline to a much broader one. It changes who works in it and how we talk about it. Last but not least, AI is not just a technological or scientific topic. It has become a social, political and geopolitical topic, which can be disorienting for such a young field and the scientists working within it, making this a particularly important moment for us to lean in and to experiment and explore.
