“AI Godmother” Fei-Fei Li Discusses the Importance of “The Voice of Humanity”!

On the afternoon of March 26th, under the dome of Gardens by the Bay in Singapore, Fei-Fei Li, a professor at Stanford University, co-director of Stanford HAI, and a world-renowned scholar in the field of AI, engaged in a face-to-face public dialogue with AI entrepreneurs in Singapore. The moderator was Professor Mohan Kankanhalli, director of the National University of Singapore’s Institute of AI. This scientist, hailed as the “AI Godmother,” used her characteristic delicate and sincere narrative to slowly unfold a conversation about technology and humanity.

When asked about the computing power race, Li Feifei calmly said, “Stanford’s lab doesn’t have a single A100 or H100.” The lab’s focus is not on how to score high, but on how to use limited resources to solve real-world problems. For instance, improving the accuracy of early cancer screening or enabling robots to understand human postures under debris during natural disasters.

● Science, not science fiction: AI must be grounded in reality rather than becoming a bubble of hype;
● Pragmatism, not ideology: Policies should balance innovation and risk, avoiding getting trapped in black-and-white debates;
● Co-construction of an ecosystem: Open-source communities, academia, and industry are all indispensable. “Walled gardens will suffocate the creativity of AI.”

It took humans 540 million years to evolve the instinct to perceive the world, while language has only a history of tens of thousands of years. However, most of the current research in AI is focused on large language models.

“This is putting the cart before the horse,” Li Feifei said directly. Her ideal foundation model must be able to capture “human perception that is more subtle than language” – the hesitation in a glance, the force of a fingertip touch, or even the silent empathy in the face of climate disasters.
● Language Intelligence: How can a symbolic system carry emotions and ethics?
● Perception-Action Intelligence: Can embodied AI truly understand the physical world like a baby through interaction with the world?
“If AI can’t learn these, it will always just be a tool, not a partner.” When the audience was struggling with the distinction between “AGI (Artificial General Intelligence) and AI”, Li Feifei shook her head frankly: “I don’t know the difference between them.”
Rather than getting bogged down in semantic disputes, she is more concerned with more practical issues –
● How can we prevent AI from exacerbating social inequality?
● How can we ensure that cleaners and programmers are equally empowered by AI?

This kind of “imperfect” sincerity, on the contrary, moved the audience. When asked if World Lab is a leading AI enterprise in the industry, she just smiled calmly: “The identity I most identify with is that of a scientist.”
She is not an entrepreneur, but just a person who is extremely curious about the things she cares about. And top scientists do not have to play the role of prophets; they only need to be sincere and ordinary people who care about the fate of humanity.
Her answer was already hidden in her research – when AI can sense the temperature of pain and the kindness of touch, it may become the lever for civilization to move towards benevolence.

Li Feifei is constantly caught between two tensions: the rationality of a scientist and the resilience of an immigrant.
● Technology ultimately comes back to “people”.
● She hopes that AI can truly solve specific problems in various industries.
● She sincerely cares about the climate crisis, social inequality, and those unseen individuals behind the system.
● When facing young students, she jokes, “Passion is more important than algorithms,” because “all journeys that follow curiosity are worthwhile.”

Amid the roaring tide of AI, the voice of humanity still deserves to be heard.