Two years before the next presidential vote, OpenAI is already trying to influence it. Chris Lehane, the company’s chief global affairs officer, has written an essay that calls the 2028 election a possible defining moment for AI. He likens the economic stakes of the technology to President Franklin D. Roosevelt’s New Deal.
Lehane’s message to politicians is that they should stop paying lip service to anti-AI sentiment and the backlash against data centers. He wants them to write policy that responds to what Americans are worried about.
That’s a big comparison to make, and at least one policy researcher rejects it.
A corporate version of the New Deal
The New Deal was a package of federal programs and reforms passed in the 1930s, when industrialization and the Great Depression had broken the old economic order. It created protections for workers and placed limits on corporate power.
According to Lehane, Americans now share a similar fear of losing control over their economic futures. He asks politicians to find a “compelling third path” so people can share in the benefits of AI. The essay sets out five policy priorities spanning worker protections, education, AI safety and data center development.
Kate Brennan, who leads AI and corporate power policy at the AI Now Institute, argued that the metaphor is a poor match for the company using it.
“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.”
She also noted that the tone departs from OpenAI’s policy agenda last year, which centered on faster AI development, fewer regulatory barriers and more infrastructure.
Published just after the midterms
The essay landed immediately after Democrats took control of Congress in the midterm elections. Their win came as voters across the country were frustrated with the cost of living.
AI itself didn’t dominate those races. Opposition to data centers did come up, and in several contests it was a major concern for both parties. Lehane could be correct that AI will claim a bigger share of the political debate over the next two years.
Affordability already looks like a central 2028 issue, as Americans contend with inflation, tariffs and rising living costs. AI may become part of that debate if voters fear data centers will raise their electricity bills, AI will take their jobs, and the economic gains from the technology will flow to a few wealthy companies.
Lots of incentives, few mandates
Lehane recognizes many of those worries. Even so, his proposals do little to prevent AI’s gains from concentrating. For the most part, he presents AI as a path to more prosperity for Americans who are willing to adopt it.
His tax code section follows that pattern. He proposes incentives for workers to learn AI skills, easier access to capital for entrepreneurs, and new models for employee ownership and profit-sharing. None of these would be binding requirements that guarantee workers a share of the wealth AI produces, or that force companies to cover the costs of displacing workers.
Many of the proposals would also push broader AI adoption and back the infrastructure build-out that companies such as OpenAI need in order to grow. The essay is much less clear on how AI companies would be held accountable for the economic disruption their products might cause.
“I was struck by how much of this policy is for a future…that is already predicated on mass AI adoption and AI benefiting society, which we have yet to see,” Brennan said. “All of it is predicated on benefiting from AI’s upside.”
Brennan said she would expect a serious plan to include a revised corporate tax rate, tax incentives for hiring people instead of automating their jobs, and measures that stop companies from gaining outsize market power.
Data centers and an Iraq war analogy
The same emphasis on opportunity over harm shows up in the data center section. Lehane acknowledges the concerns, then turns to the construction jobs and local tax revenue the facilities could generate.
He also goes a step further. Politicians who support a data center moratorium, he warns, could endanger America’s lead over China, and he likens that stance to the 2002 vote authorizing the Iraq war.
Children, schools and an untested assumption
Child safety and education are close to the core of the agenda. Lehane puts weight on age-appropriate protections and on safeguards for minors using AI companions.
He also opposes keeping AI out of classrooms, arguing that doing so could open a gap between wealthy students and “children in rural communities.” That position depends on AI improving learning, which hasn’t been proven. Critics say the claim should face more scrutiny than proposals for extra funding or more one-on-one time with human teachers.
Safety pledges and a history of pushback
Lehane calls for enforceable AI safety standards, independent evaluations and public oversight. He also wants the ability to restrict AI systems or switch them off with a kill switch. Recent studies show that OpenAI and other AI labs have not published complete plans for what they would do if a model escaped human control.
Brennan described those positions as “disingenuous,” citing OpenAI’s record of opposing legislation designed to require those same measures.
In 2024, OpenAI came out against California’s AI safety bill, SB 1047, which would have required catastrophic risk evaluations and shutdown capabilities. Last year the company also sought to limit state regulation, claiming states could undermine American innovation.
Some of those positions have since shifted. OpenAI has supported several state safety laws and said it would be open to having independent evaluators embedded inside the company.
The question of control
Lehane further argues that people should retain control over consequential decisions made with AI, without handing their judgment entirely to machines. What that control would look like is hard to pin down when people are increasingly required to use AI in education, work, healthcare and government.
Throughout the essay, Lehane casts AI as a way to restore Americans’ sense of agency and economic opportunity. He also ties the “middle class” success of ordinary Americans to participation in an economy that companies like OpenAI are increasingly shaping. The essay leaves open whether his policies would distribute power more widely or quietly entrench it for the companies drafting them.