OpenAI is already trying to shape the 2028 election

Chris Lehane, OpenAI's chief global affairs officer, published an essay arguing that 2028 could define AI the way the New Deal redefined the economy after industrialization and the Depression. He wants politicians to stop empty anti-AI gestures and build a "compelling third path" so Americans share the upside.

The timing is deliberate. Democrats just flipped Congress in the midterms on cost-of-living anger. Data-center opposition showed up as a bipartisan issue in several races. Affordability — bills, jobs, who captures the gains — is already shaping the next presidential cycle. Lehane is putting OpenAI's preferred frame on that conversation early.

Kate Brennan of the AI Now Institute told TechCrunch this is a notable shift from last year's accelerate-and-deregulate posture. The question is whether the new language redistributes power or just rebrands the same expansion agenda.

We're glad a frontier lab is finally talking worker protections, education, child safety, and enforceable safety standards in the same breath as infrastructure. We are not buying the New Deal metaphor without the New Deal's constraints on private power.

Lehane's five priorities span workers and education, AI safety, and data-center development. Dig into the tax ideas and you get incentives to learn AI skills, easier capital for entrepreneurs, and softer talk about employee ownership — not binding rules that workers share AI-generated wealth or that firms pay for displacement. Brennan's critique is the one we're carrying: invoking state capacity to benefit the public, written by a private company advocating for itself, is "ahistorical at best and harmful at worst."

On data centers, Lehane nods at local concerns, highlights construction jobs and tax revenue, then warns that moratorium-friendly politicians could cost America its lead over China — even comparing that stance to the 2002 Iraq War authorization vote. That is not bridge-building. That is moral pressure dressed as history.

Child safety and classroom AI get central billing. Age-appropriate protections around AI companions for minors are overdue and worth supporting. The claim that keeping AI out of classrooms deepens inequality between wealthy and rural students assumes AI improves learning — a premise that remains unproven and that TechCrunch rightly flags.

On safety, Lehane wants enforceable standards, independent evaluations, oversight, and a kill switch. Fine words. Brennan calls them "disingenuous" given OpenAI's 2024 opposition to California's SB 1047 and last year's push to limit state rules. The company has since softened toward some state safety laws and says it is open to embedded independent evaluators. I'm watching deeds, not essays.

Useful AI deserves a political coalition that is not only fear or only hype. Lehane is right that lip service helps nobody. He is wrong if "third path" means adopt first, redistribute later. Agency for the middle class cannot be defined as participation in an economy shaped primarily by firms like OpenAI.

TechCrunch reports OpenAI did not respond to questions seeking more specifics before publication. AI was not a dominant midterm issue overall, but data-center fights were. Lehane's essay arrives as inflation, tariffs, and living costs dominate the early 2028 conversation — terrain where electricity bills and job anxiety will keep pulling AI into politics.