OpenAI Staffers Quietly Back Rival Super PAC Challenging the Industry’s AI Push

Several current and former OpenAI employees have given more than $215,000 to Guardrails Alliance, a new super PAC that wants tougher rules for frontier AI companies and is positioning itself against a far better funded industry-backed political operation. The donations, which came as OpenAI leadership faces questions over its ties to pro-industry spending, underscore a widening internal split over how aggressively AI should be regulated in the 2026 election cycle.

Guardrails Alliance says it has launched with $5 million in initial support and hopes to raise $15 million this cycle, placing it in direct political opposition to Leading the Future, a pro-AI super PAC with more than $100 million behind it and support from major technology figures, including OpenAI cofounder and president Greg Brockman. The clash is turning OpenAI employees into unlikely participants in the fight over who gets to shape the rules for the technology their company is helping build.

Why OpenAI employees are funding a rival super PAC

The simplest explanation is that some OpenAI workers believe the company’s political allies are pushing the AI industry too far toward self-regulation and too far away from binding safeguards. Their donations to Guardrails Alliance amount to an internal rebuke of Leading the Future, the political machine that has emerged around the industry’s argument for lighter-touch oversight.

WIRED learned that seven current OpenAI employees and one former employee have contributed to Guardrails Alliance. The group shared some donor names ahead of its first quarterly filing with the Federal Election Commission, which is set to become public on July 15. Two OpenAI employees are expected to appear in that filing, while five additional staffers will be listed in later disclosures.

Although the employee donations are small relative to the enormous sums flowing through AI political action committees, they are notable because they come from people inside one of the world’s most influential AI companies. In effect, the money signals that not everyone at OpenAI is comfortable watching the company’s broader network move toward an industry-first political strategy.

What the donations say about internal tensions at OpenAI

The contributions suggest that the debate over AI policy is no longer taking place only between companies, lobbyists, and lawmakers. It is also happening inside the organizations building frontier systems, where researchers and engineers are increasingly asking whether the public will receive enough protection as commercialization accelerates.

OpenAI has long framed safety work as central to its mission, but employees contributing to Guardrails Alliance appear to be arguing that research alone is not enough. They want those concerns translated into legal requirements and election spending that can counterbalance a growing anti-regulation push.

That internal unease has sharpened around the political activity of Brockman, who has committed substantial personal funds to Leading the Future. Some staffers have pressed OpenAI leadership for clarity about how the company relates to the super PAC, prompting the company to stress that these efforts are personal and not corporate policy.

“Employees are free to participate in the political process in their personal capacities,” OpenAI has said, adding that this includes donating to candidates, campaigns and political organizations.

Who is paying into Guardrails Alliance?

One of the largest known gifts came from Juan Felipe Cerón Uribe, a research engineer at OpenAI since 2022, who contributed $200,000. That single donation accounts for the overwhelming majority of the OpenAI-linked money identified so far.

Cerón Uribe said he has spent four years working on ways to reduce the possible harms that AI could cause at scale. In his view, that work only matters if it leads to concrete guardrails that prevent private companies from operating without accountability.

Cerón Uribe said he worried that years of safety research could be wasted unless it results in guardrails that hold private companies responsible for how they develop AI.

He also argued that ultra-wealthy technology backers are trying to keep the sector unregulated, and said that backing Guardrails Alliance was an easy decision once he saw the group challenging that effort.

Other current OpenAI staffers who donated include Gabriel Wu, an OpenAI safety researcher, who gave $5,000, and alignment researchers Julie Steele and Jason Wolfe, who each gave $5,000 as well, according to the super PAC. Former OpenAI research manager David Farhi, who left last summer after seven years at the company, contributed $3,000 and will appear in the group’s July filing.

Farhi described AI as a technology with exceptional upside and serious risks, saying his disappointment stems from the way pro-industry forces have worked to narrow the conversation before regulation has a chance to take shape.

Known OpenAI-linked donations to Guardrails Alliance

How big is the fight over AI political spending?

The money flowing into competing AI super PACs has grown quickly enough to reshape the 2026 political landscape before voters even begin focusing on the race. Guardrails Alliance is aiming for $15 million, but that target is dwarfed by Leading the Future’s reported war chest of more than $100 million.

Leading the Future has secured support from tech leaders such as Brockman and has positioned itself as a force that will challenge candidates and policies it sees as hostile to innovation. Brockman and his wife Anna have pledged $50 million to that effort.

Guardrails Alliance, by contrast, is trying to sell a different idea: that AI policy should be driven by workers, unions and public-interest groups rather than only by corporate donors. Its backers say that contrast is politically valuable even if they cannot match the other side’s spending power.

Guardrails Alliance cofounder Shaunna Thomas said the group does not need to spend dollar for dollar with its opponents, arguing that public exposure of the AI super PACs can be more effective and less expensive than an arms race.

Thomas, a veteran Democratic political organizer, said the existence of a $100 million threat hanging over lawmakers makes it harder for politicians to advance tougher AI rules. Her answer is to build a coalition that can follow Leading the Future into contested races and make the issue visible to voters.

Why the financial imbalance may not decide the outcome

On paper, the pro-industry side has a commanding edge. In practice, political spending does not always translate neatly into influence, especially when the underlying message is unpopular or perceived as too close to wealthy insiders.

Guardrails Alliance believes it can exploit that gap. Rather than compete solely on ad spending, it plans to frame AI regulation as a public accountability issue and treat the super PAC ecosystem itself as part of the story.

That approach could resonate with labor organizations, watchdog groups and voters who see advanced AI systems as too powerful to be left to private companies alone.

What does Leading the Future want?

Leading the Future says it exists to stop policies that would choke off innovation and to back candidates who share that view. In its public messaging, the group presents itself as a constructive political force that wants proactive federal rules, but only ones that do not burden the industry.

The group has already shown a willingness to intervene directly in races. Among its early efforts was a campaign to defeat Alex Bores, the New York state lawmaker who wrote the state’s landmark AI safety legislation and later ran for Congress. Bores lost in a primary last month.

Leading the Future has since backed a range of pro-industry candidates across the country. A spokesperson for the group rejected the idea that it is trying to shut down debate, saying it supports a clear policy agenda and has endorsed a broad mix of candidates and officeholders.

The group said it has not sought to suppress public discussion about AI and pointed to its own previous support for federal regulations.

OpenAI’s global affairs chief, Chris Lehane, has previously said he helped set up Leading the Future and has occasionally advised Brockman on political giving, though he is not involved in day-to-day PAC operations.

Why this matters for OpenAI and the AI policy debate

The OpenAI employee donations matter because they expose a disconnect between the public posture of an AI company and the political instincts of some of the people building its systems. They also show how quickly AI policy has become a high-stakes electoral issue rather than a purely technical or academic debate.

For years, discussions about AI safety often centered on research papers, internal red-teaming, and voluntary best practices. Now, as frontier systems become more capable and more commercially important, the fight has moved into campaign finance, where competing super PACs are trying to define the rules before Congress or state legislatures do.

That shift has several implications:

• Frontier AI policy is becoming a major election issue in 2026.
• Employees inside leading AI firms are using personal money to influence that debate.
• Super PAC spending is increasingly shaping which lawmakers can safely support regulation.
• Calls for AI safeguards are now competing directly with industry-backed anti-regulation campaigns.

How other AI-related PACs fit into the picture

Guardrails Alliance is not the only political group trying to counter industry spending. Public First Action, a super PAC backed by Anthropic with $20 million, has also said it will promote AI safeguards and challenge pro-industry messaging in the 2026 elections.

That means the political fight is not a simple one-on-one contest between a single corporate camp and a single reform camp. Instead, multiple AI companies, policy advocates and political strategists are creating overlapping coalitions around different versions of the future: one that prioritizes rapid deployment, and another that insists on stronger external oversight.

Both Guardrails Alliance and Public First Action backed Bores in his congressional primary, a contest that drew $27 million in total spending from the pro-AI and pro-safeguard camps combined. The size of that number suggests AI policy is no longer a niche issue; it is a serious source of campaign spending and strategic coordination.

What comes next in the 2026 election cycle?

Guardrails Alliance is expected to reveal more donors in its filing, including former Andreessen Horowitz partner John O’Farrell, though his exact contribution has not been made public. O’Farrell did not respond to requests for comment, but he has already criticized his former colleagues for allegedly using Leading the Future to pressure politicians away from aggressive AI oversight.

The super PAC says it is also looking at additional Democrats in 2026, including candidates in California’s 34th congressional district. That suggests the campaign may broaden beyond New York and into other races where AI regulation could become a distinct wedge issue.

For now, the broader significance is not the amount one group or another has raised. It is the fact that people inside the company most associated with frontier AI are spending their own money to oppose a PAC backed by one of OpenAI’s best-known leaders.

That is a strong sign that the politics of AI are no longer being defined only by executives, investors and Washington strategists. The engineers and researchers working closest to the technology are starting to make their own voices heard, and they are doing so through the hard-edged machinery of election spending.

Key timeline

The growing split inside OpenAI is unlikely to disappear soon. As AI systems become more central to politics, business and public life, the arguments over how much control private companies should have will only intensify. For some OpenAI employees, the answer is to spend their own money to ensure the next generation of AI laws is written with stronger guardrails from the start.