Can AI Be Safe in a System Built to Race?

ABOUT THIS EPISODE: What happens when the economic incentives driving AI development come into conflict with the obligation to protect people? In this conversation, Muxin Li unpacks the race dynamics behind the AI boom: why companies and governments are rewarded for moving faster, scaling bigger and staying ahead, even when the social consequences remain unresolved. We talk about the costs that are rarely borne by the companies making these decisions: labor displacement, loss of consent, rising demands on water and electricity, pressure on local communities, and the erosion of public control over how AI infrastructure is developed and deployed. We also examine the promise of a coming "golden age for everyone," and ask whether that vision reflects broad public benefit or the interests of investors and industry. The deeper question is whether responsible AI can emerge from a system in which the strongest incentives favor growth, competition and return on investment. What would have to change for public protection to carry comparable weight? Who should have a say beyond regulators and industry insiders, and what kind of society should AI ultimately serve?## ABOUT THIS EPISODE AI companies are rewarded for moving faster, scaling faster and beating competitors to the next breakthrough. What happens when safety asks them to slow down? Juliet MacDowell speaks with Muxin Li, product leader and founder of Human Inference, about the economic and political systems shaping AI development. They examine the incentives driving the AI race, what society is contributing to it, who captures the value, and what it would take to distribute power and benefits more broadly. The CENTRAL QUESTION Can AI be developed safely inside a system that rewards the race? KEY TOPICS Why companies continue racing toward increasingly powerful AI despite acknowledged risks How fear, rivalry and status shape decisions about AI development What society contributes through labor, data, public resources and local infrastructure Who captures the value created by the AI boom How democratic participation could provide a counterweight to concentrated power Whether different measures of progress could change what AI development rewards Why THIS CONVERSATION MATTERS AI safety is often approached as a problem of models, safeguards and testing. But the systems building AI also respond to competition, investment, political power and expectations of growth. This conversation looks at what happens when those incentives become part of the safety question itself, and asks what kind of institutions could give the public a greater role in deciding how AI develops and how its benefits are shared. IN THIS EPISODE *00:00* When AI safety asks companies to slow down *00:46* Technical uncertainty and the race to build anyway *02:26* Why race toward superintelligent AI? *03:12* Fear, rivalry and status competition *04:07* What society is contributing to the AI boom *06:20* The promise of an AI-driven “golden age” *07:27* Public buy-in, geopolitics and who gets the benefits *09:09* Should data centers provide their own power and water? *11:02* What can counter concentrated AI power? *11:56* Bringing the public into AI governance *14:27* Can AI make democracy more participatory? *16:04* Corporate incentives and short-term thinking *20:24* Measuring AI's impact on society *22:27* Sharing the benefits of AI more broadly *24:21* What would a society where everyone wins look like? *26:31* Why speed keeps winning *26:57* Outro The Guest Muxin Li** is a product leader and founder of Human Inference. Her work examines the economic and political systems behind technology, with a particular interest in how power is distributed and how technology can strengthen civic participation. She is developing tools that make political power more visible and help citizens understand how their representatives vote and who funds them. Her broader work explores how technology, political institutions and economic incentives might be designed to give people greater agency over the systems that shape their lives. BETWEEN US AND THE MACHINE Between Us and the Machine examines how AI is reshaping power, governance and human agency, and the choices that will determine its place in society.