Kate Crawford

Some people build AI and others criticize it from the outside. Kate Crawford does something more uncomfortable: she studies it from within, using data, academic tools, and scientific rigor, but asking questions the tech sector preferred to avoid. Born in Australia in 1974, she is a researcher, writer, and professor who has spent more than two decades focused on one core question: what systems of power underpin artificial intelligence?

The work that put her on the global map arrived in 2021 with "Atlas of AI: Power, Politics and the Planetary Costs of Artificial Intelligence" (Yale University Press). The book does not talk about robots or apocalypse; it talks about something more concrete and more urgent: the minerals extracted to manufacture chips, the data annotation platform workers earning cents per hour in the Global South, the energy consumed by data centers, and how all of that is part of a system usually presented as clean and virtual. "Atlas of AI" was named a best book of the year by the Financial Times, won three international prizes, and was translated into twelve languages. It is not a technical text: it is a political map of something we thought we understood.

Long before that book, Crawford had already made her mark. In 2017 she co-founded the AI Now Institute at New York University, the first research center dedicated exclusively to the social impacts of artificial intelligence, including bias, surveillance, and labor rights. She also co-founded the FATE project (Fairness, Accountability, Transparency, and Ethics in AI) at Microsoft Research, where she is a senior principal researcher. In 2023 she was named among TIME magazine's 100 Most Influential People in AI. Today she is a research professor at USC Annenberg and holds the inaugural visiting chair for AI and Justice at the École Normale Supérieure in Paris.

What I find most valuable about Crawford is not that she is critical, but that she is specific. She does not say "AI is bad": she says "here is the mine in the Congo that produces the cobalt for your phone, here is the worker in Kenya reviewing violent content to train the model." That specificity is what makes her work difficult to ignore, even for those who disagree with her. If you use AI every day (and you probably do), knowing where that AI comes from, the real cost of each response, is a way of using the tool with your eyes open. And that, precisely, is what Crawford has spent two decades trying to give us.