Ezra Klein, We Have to Take the Future of A.I. Into Our Own Hands, NYTimes, May 31. 2026. Here's Claude's summary of the article:
Ezra Klein's piece argues that the public conversation about AI has been almost entirely focused on preventing harms, while largely ignoring a different question: how can AI actively benefit the public?
He acknowledges the skeptical climate — polling is poor, data centers are controversial, and Pope Leo XIV's first encyclical addressed AI's dangers — but insists that since AI is here regardless, the meaningful question is how it gets directed and for whom.
His core argument is that AI's public benefits won't happen automatically. They require deliberate investment in data, financing, and compute. He points to impressive examples of AI solving hard scientific problems (protein folding, pancreatic cancer detection, a new pulmonary fibrosis drug, weather prediction) to show what's possible when it's properly deployed — but notes these successes depended on pre-existing public infrastructure like the Protein Data Bank, itself publicly funded since the 1970s.
His proposed agenda includes: a publicly controlled frontier-level AI model; subsidized compute access for universities and public agencies (to close the widening gap between institutions like Goldman Sachs and public institutions); government-guaranteed markets for AI-driven solutions to public problems (analogous to Operation Warp Speed); using AI to help citizens navigate government services; and funding the creation of clean, novel public datasets.
He closes by noting that we've been consumed by what we fear AI will do to us, at the expense of asking what we hope it will do for us.
