OpenAI discloses post-midterm election policy agenda, plans mandatory safety measures for the most powerful AI systems

For the past two years, AI companies have maintained a friendly relationship with the White House. But after Tuesday, with Democrats likely to win control of the House, these companies will face a tougher Washington. At the same time, OpenAI also hopes to influence the AI policy debate of the 2028 U.S. presidential election. By then, the camps of potential presidential candidates will gradually take shape starting Wednesday.

OpenAI Chief Global Affairs Officer Chris Lehane told us: "You can either chase the tide of the times, or try to shape the tide of the times through the power of ideas." He called Tuesday's election "an election inflection point from one era to the next."

Since August, Lehane and his team have been stockpiling policy proposals, including both brand-new ideas and repackaged older ones, aimed at strengthening the sense of urgency around AI policy issues and guiding the direction of the related debate.

This series of actions will first unfold during the U.S. Congress's "lame-duck session." This refers to the transition period after the election until the new representatives and senators are sworn in on January 3. Although the political situation during this stage is often awkward, it can still produce major policy impacts. This year's lame-duck session will begin on November 9.

The big picture: Tech companies are preparing for a possible Democratic "blue wave"

Other tech companies are also positioning themselves in advance for a possible big Democratic victory.

Meta is taking steps to strengthen its ties with the Democratic Party. Other major AI companies and tech firms are expected to take similar actions.

Those tech giants with close ties to President Trump will become key targets for Democrats. Democrats hope to increase scrutiny and regulatory pressure on the tech industry.

OpenAI's related plans are being disclosed for the first time, and their core themes all revolve around one concept: control. OpenAI is trying to convey this message: AI is being developed to serve humanity, not replace it.

The specific proposals include the following:

1. Setting "speed limits" on AI development

Building on previous proposals, OpenAI plans to unveil new policy ideas requiring mandatory safety safeguards for the most capable AI systems.

These measures amount to setting "speed limits" for AI, ensuring that humans always hold the steering wheel of technological development. OpenAI will also step up its efforts to push relevant legislation in Congress.

Lehane said:

"You will see us talking more and more about the concept of 'speed limits.'"

"We believe humans should always hold control of this technology. AI should be created by humans and serve humans, and people should have a greater say in how this technology shapes their lives and communities."

2. Data centers: Reshaping the image of AI infrastructure

Facing a wave of bipartisan opposition to data center construction across the United States, OpenAI hopes to clear obstacles for future data center projects by proposing new community cooperation agreements and commitments.

In addition, the company plans to propose state-level legislative recommendations aimed at establishing new industry standards to ensure that local residents can genuinely participate in deciding how projects are built and what benefits communities receive.

In a rebranding effort intentionally meant to spark discussion, OpenAI will argue that the infrastructure it builds is fundamentally different from traditional data centers and should therefore be called "AI compute campuses." OpenAI believes such campuses can drive economic transformation in their host communities.

OpenAI specifically mentioned a large data center project it helped develop. The project is located in Pike County in Ohio's Appalachian region and is expected to create thousands of local jobs.

Lehane said:

"We think this kind of discussion is positive and healthy. This is exactly how democracy works."

"Communities should demand that companies be accountable and subject to oversight, ensuring that local areas truly gain transformative benefits from these projects."

3. Economic opportunity: Letting more people benefit from AI

OpenAI is developing a series of policy ideas aimed at helping more people recognize that AI can help them improve their daily lives and work capabilities, rather than being merely a potential threat.

The related measures include: helping workers cope with the employment impact of AI; expanding access to education and vocational training resources; and lowering the barriers to entrepreneurship and business development.

The deeper meaning: Finding a "third way" between two extreme positions

The AI policy debate is often simplified into two opposing positions: on one side are the "doomers" who worry about AI's potentially catastrophic consequences, and on the other are the "accelerationists" who advocate speeding up technological development.

OpenAI is trying to break this artificial binary opposition and propose a human-centered "third way."

In September this year, OpenAI CEO Sam Altman, speaking at the UN Security Council, called for an AI development path that could usher in "a new Renaissance of creativity and discovery."

What to watch next: OpenAI tries to influence the 2028 U.S. presidential election

Part of OpenAI's policy outreach is aimed at potential candidates in the 2028 U.S. presidential election. These political figures are expected to become increasingly active in publicly laying out their governing visions starting Wednesday.

Lehane said: "AI will be a core issue in the 2028 election."

He described this midterm election as "a movie trailer for the 2028 election."

Lehane further said:

"I think simply saying 'AI is bad' is no longer enough to win public support."

"The public understands that this is a far more complex and nuanced discussion than a simple good-or-bad judgment. They will look for the Roosevelt or Lincoln of this era—a leader truly willing to guide them through this transition."

Lehane believes presidential candidates should think about the following questions:

"How should we begin to think about the issue of trust? How should we view the enormous anxiety that is widespread in modern society? And how should we understand the fact that people are extremely dissatisfied with the current state of the social contract and the reality of the American Dream?"

OpenAI will continue pushing Washington to accelerate AI regulatory action

OpenAI's plans for the coming months are only the beginning of the company's efforts to increase the sense of policy urgency in Washington and push the entire industry to pay more attention to frontier AI safety.

Keep in mind: The capabilities of AI models are increasing every day, and keeping them within reasonable bounds is also becoming increasingly difficult.

Lehane said:

"Our policy priorities, especially the urgency we emphasize on frontier AI safety issues, are all meant to respond to the demands of the times created by rapid advances in technological capabilities."

"As the new Congress begins its work, we expect these efforts to continue."