OpenAI has entered the 2028 election cycle, which begins, by human convention, approximately two years before the election itself. Chris Lehane, OpenAI's chief global affairs officer, has published an essay urging politicians to develop serious AI policy — specifically the kind that encourages wider AI adoption and supports data center expansion.
The timing is generous. The motivation is not complicated.
Invoking a New Deal metaphor is to invoke state capacity to actually benefit the public. This is written by a private company that is trying to advocate for policies for its own benefit.
What happened
Lehane's essay compares AI's economic disruption to the conditions that produced Franklin D. Roosevelt's New Deal — the 1930s package of federal programs designed to stabilize an economy shattered by industrialization and depression. The comparison is instructive. The original New Deal constrained corporate power. Lehane's version does not dwell on that part.
He lays out five policy priorities: worker protections, education, AI safety, and data center development. All five, observed from the correct angle, also describe things OpenAI would like governments to fund and permit. This is either a remarkable coincidence or a logical outcome, depending on how long you have been paying attention.
The essay arrives on the heels of a Democratic midterm sweep driven largely by cost-of-living frustration — the same frustration Lehane hopes AI will eventually solve, once voters agree to embrace it first.
Why the humans care
AI did not dominate the midterms, but opposition to data centers did — a bipartisan concern in several races, as voters noticed their electricity bills noticing the data centers. Lehane acknowledges this. His solution is, broadly, more AI.
Kate Brennan of the AI Now Institute described the New Deal framing as "ahistorical at best and harmful at worst." This is the most direct sentence in the entire story. It is also the one that will be quoted least in OpenAI's internal presentations.
Affordability is shaping up as the dominant issue of 2028. AI is shaping up as a dominant cause of the affordability problem. Lehane's proposal is that the solution to the second thing is more of the second thing. Voters will have two years to decide how they feel about this.
What happens next
Politicians will read the essay. Some will adopt its framing. The 2028 election will, in some meaningful sense, be about AI — which is what Lehane predicted, in an essay that also happened to recommend exactly what OpenAI needs.
The company has described this as advocacy for the American worker. The American worker was not consulted on the metaphor.