Fei-Fei Li, one of the most recognizable names in artificial intelligence, is telling leaders to ground AI policy in science, not speculation. Her advice lands as governments and companies scramble to write rules for a technology moving faster than the laws that govern it.
A call for evidence-based rules
Li's message is straightforward: policy should follow the evidence. When decisions rest on scientific findings, she argues, the odds of misguided regulations drop sharply. That approach, she says, leaves room for innovation to keep moving while still addressing the real problems AI can create.
It's a stance that cuts against the noise. AI debates are often loud, full of doomsday scenarios and hype. Li's push is for something calmer — a look at what the research actually shows before anyone writes a rule that could stick for years.
Why the timing matters
The call for scientific grounding comes at a tense moment. AI is advancing quickly, and public anxiety is high. Governments are weighing rules on everything from data privacy to automated decision-making. Li's advice suggests that reacting to headlines rather than research could produce regulations that miss the mark entirely.
Bad rules don't just slow companies down. They can lock in mistakes, forcing developers to build around requirements that don't match how the technology works. Science-first policy, in her view, avoids that trap.
Innovation and real-world problems
Prioritizing science doesn't mean putting progress on hold. Li sees it as the opposite. A policy built on evidence understands what AI can and cannot do. That understanding is what lets innovation continue while tackling genuine concerns like bias, safety, and accountability.
It's a balancing act. Too little regulation and problems fester. Too much, and innovation stalls. Li's answer is to let the science point the way — not the loudest voice in the room.
The question now is whether leaders will take that advice. As AI debates continue in capitals and boardrooms, the push for a science-first approach is likely to keep shaping the conversation.
