OpenAI prepares its agenda for the 2028 election: an American-style 'New Deal'

Lehane's essay: five priorities and a Roosevelt parallel

Chris Lehane, head of global affairs at OpenAI, believes the 2028 presidential election could be a defining moment for artificial intelligence, and he calls on politicians to develop policies in advance that address Americans' concerns, rather than simply exploiting anti-AI sentiment. In the published essay , he compares AI's potential to transform the economy with Franklin Roosevelt's New Deal — the package of federal programs of the 1930s — noting that Americans today feel a similar anxiety about losing control over their economic future, and he calls for finding a 'compelling third way' that would allow people to share in AI's benefits. He lays out his agenda in five priorities: protecting workers, education, AI safety, and data center development.

Political context and criticism

The essay appeared shortly after the Democrats' major victory in the midterm elections, when the party gained control of Congress amid voter discontent over the cost of living. Although AI was not the dominant issue, opposition to data center construction became a notable bipartisan issue in a number of campaigns. Kate Brennan, who leads the AI policy and corporate power work at the AI Now Institute, told TechCrunch that Lehane's rhetoric marks a noticeable shift compared with OpenAI's agenda from last year, which focused on accelerating AI development, lowering regulatory barriers, and expanding infrastructure. In her words, 'invoking the New Deal metaphor is an appeal to the state's ability to genuinely serve society,' whereas the document was written by a private company advancing its own interests, which is 'at best ahistorical, at worst harmful.' TechCrunch also notes that many of Lehane's proposals encourage broader AI adoption but contain no binding requirements that would guarantee workers a share of the AI-generated wealth or compensation to companies for displacing jobs.

Data centers, safety, and contradictions with the previous position

On data centers, Lehane acknowledges voters' concerns — including rising electricity bills — but emphasizes jobs and local tax revenue, and he warns politicians who support construction moratoriums about the risk to US leadership over China, comparing such a stance to the 2002 vote authorizing the war in Iraq. On safety, he advocates for enforceable standards, independent audits, public oversight, and the ability to 'emergency shut down' systems; however, as TechCrunch notes citing recent research, OpenAI and other labs have not published comprehensive plans of action for cases where a model goes out of control. Brennan called these positions 'insincere,' recalling that in 2024 OpenAI opposed the California law SB 1047, which required assessments of catastrophic risks and shutdown mechanisms, and last year pushed to limit regulation at the state level. At the same time, the company has recently adjusted some of its positions, supporting several state safety laws and stating its willingness to admit independent auditors into the company. Separately, Lehane defends child safety and access to AI in schools, warning that rejecting the technology would deepen the gap between affluent students and children from rural communities — although, as TechCrunch points out, the claim that AI improves learning remains unproven. OpenAI did not respond to TechCrunch's request for details on the proposals by the time of publication.