OpenAI’s Lehane likens backing a data centre ban to the 2002 Iraq vote

AI will be the number-one issue of the 2028 US presidential election, says Chris Lehane, OpenAI’s chief global affairs officer. His memo, titled “2028: The AI Election”, sets out five questions he says candidates must answer to build an AI agenda.

Politico published the Chris Lehane memo in full on Monday, a day before the US midterm elections.

Lehane writes that 2028 has echoes of 1932, the last peacetime presidential race held during a change on this scale. Franklin D. Roosevelt won it, and his New Deal reimagined the social contract.

A candidate who sets out a “compelling third path” will have a chance to define a new contract, Lehane writes. That path runs between those who would let AI advance “without speed limits” and those who would keep control over the technology “in the hands of a self-appointed elite”.

The data centre warning

By early October, one in five ads by Democratic candidates or groups mentioned data centres, according to the memo. Of the Democratic ads that criticised data centres in October, the vast majority ran in a handful of states, the memo says. They were mainly Michigan, Ohio, Pennsylvania and Texas, where the sites are being planned and built.

At an early 2028 Democratic event, Lehane writes, candidates could be asked to raise their hands to back a moratorium on data centres. Six months later, the US could lose its AI lead to China for lack of computing power, he adds.

“This could well be the 2028 version of the 2002 vote to go to war in Iraq,” Lehane wrote.

He also writes that OpenAI’s analysts found likely Chinese Communist Party influence operations in the US. They posted social media comments and images claiming that AI buildouts were raising household electricity bills, according to the memo.

Polls and five questions

The memo cites polling that Lehane says he has seen or run himself. In his own polling, voters prefer a candidate who leads their party in shaping AI policy over one who follows it, by 25 points. Some 66% support continued AI development, including 53% who want stronger safeguards, while 24% favour a pause.

The five questions cover safety and human control, taxes, children and jobs, affordability, and the race with China. On safety, Lehane writes that “AI is a machine, not a person”. He calls for enforceable safety standards, independent assessments and the power to restrict or shut down systems when needed.

On jobs, he writes that OpenAI’s compute campus in Piketon, Ohio alone will need 25,000 skilled tradespeople.

A push after the midterms

OpenAI will launch a policy push right after Tuesday’s vote, Jim VandeHei and Mike Allen reported for Axios on Sunday. It starts in the lame-duck session of Congress, which opens on 9 November.

OpenAI will unveil new ideas for mandatory safeguards on the most powerful AI systems, which it calls “speed limits”, Axios reported. In September, OpenAI asked Congress for national safety rules. It will also offer community deals for new data centres. And it will propose state laws that give residents a say in how the sites get built.

“You’re going to see us talk more and more about the concept of speed limits,” Lehane told Axios.

OpenAI also wants its data centres called “compute campuses”. A spokesperson confirmed the new name to Business Insider.

“Compute campuses better describes the scale and role of the infrastructure we’re building for AI,” the spokesperson said.

Last week, TNW reported that voters in Iowa and Ohio back AI safety measures over winning the AI race. Lehane’s memo itself says there was “little room to be ‘for’ AI infrastructure” in the 2026 campaigns.

For the midterms, OpenAI and Anthropic will each run a “war room”, Ashley Gold reported for Semafor. Their job is to keep election information on the two companies’ platforms accurate. It is a first for both.