OpenAI has started working on the 2028 presidential election, and voters won’t pick a nominee for almost two years. According to TechCrunch AI, chief global affairs officer Chris Lehane published an essay comparing AI’s economic potential to Franklin D. Roosevelt’s New Deal. He urges politicians to find a “compelling third path” that lets Americans share in AI’s benefits.
The timing matters. The essay landed right after Democrats flipped Congress in midterms shaped by frustration over the cost of living. OpenAI wants to set the terms of the AI debate before candidates do, and a frame that sticks early in a campaign is hard to move.
What OpenAI is asking for
Lehane lays out five policy priorities covering worker protections, education, AI safety and data centers. Here’s how some of them read:
- Taxes: Incentives for workers to learn AI skills, easier capital access for entrepreneurs, and new approaches to employee ownership and profit-sharing.
- Data centers: He acknowledges concerns, then points to construction jobs and local tax revenue.
- Kids and schools: He backs age-appropriate protections and safeguards around AI companions, but warns against keeping AI out of classrooms.
- Safety: He calls for enforceable standards, independent evaluations, public oversight and the ability to restrict systems or shut them down with a kill switch.
The sharpest line is about data centers. Lehane warns that backing a moratorium could jeopardize America’s lead over China, and he compares that stance to the 2002 vote authorizing the Iraq war. That’s a pointed message to ambitious politicians: this vote will follow you around. It also suggests how seriously OpenAI takes the backlash.
The case against the New Deal comparison
Kate Brennan, who leads AI and corporate power policy at the AI Now Institute, told TechCrunch that Lehane’s rhetoric is a notable shift from OpenAI’s agenda last year, which focused on speeding up development and cutting regulatory barriers. She isn’t convinced by the new framing.
“Invoking a New Deal metaphor is to invoke state capacity to actually benefit the public,” Brennan said. “This is written by a private company that is trying to advocate for policies for its own benefit. It is ahistorical at best and harmful at worst.”
I think that’s the weakest point in the essay. The original New Deal protected workers and put limits on corporate power. Lehane’s tax ideas are mostly incentives, with no binding requirement that workers share in AI wealth or that companies pay for displacement. Brennan would rather see a revised corporate tax rate, tax incentives for hiring humans over automating, and limits on market power.
She also called OpenAI’s safety stances “disingenuous.” The company opposed California’s SB 1047 in 2024, a bill that would have mandated catastrophic risk evaluations and shutdown capabilities, and it later sought to limit state regulation. In fairness, OpenAI has since backed several state safety laws and says it’s open to independent evaluators embedded in the company.
Where this goes by 2028
Lehane’s prediction that 2028 could be a defining moment for AI deserves some skepticism, since he’s an advocate with a company to protect. The evidence still gives him partial support. AI wasn’t a dominant midterm issue, but data center opposition became a major bipartisan concern in several races, and affordability is shaping up as a central 2028 theme.
Here’s my read on the next two years:
- 2027: Electricity bills become the main way AI enters local politics, with more fights at the state and county level.
- Early 2028: Primary candidates get pushed to take a position on data center moratoriums. Since the opposition is bipartisan, neither party gets a clean answer.
- Late 2028: The argument moves to voluntary incentives versus binding rules, meaning profit-sharing perks on one side and taxes or displacement costs on the other.
What to do now
- If you build or lease data centers: Treat community opposition as project risk, and have a credible answer on who pays for grid costs.
- If you’re rolling out AI at work: Track what happens to jobs and retraining now. You’ll want that record if displacement rules or hiring incentives gain traction.
- If you sell AI to schools or families: Collect evidence that it improves learning. TechCrunch notes that the claim remains unproven, and age safeguards are becoming the baseline.
- If you run compliance or policy: Plan for state rules. And compare what your vendors say in public with what they lobby for.
The real test of Lehane’s essay comes when a binding rule would cost OpenAI money. Watch how the company responds to the next state bill with teeth.
The full report is at TechCrunch AI.