There is now an open political war over who gets to write the rules for AI, and it is playing out in real elections with real money.
The Guardrails Alliance, launched this week, is a super PAC, a political fundraising group that can raise and spend unlimited money on ads and campaigns as long as it does not hand money directly to candidates. It was founded by Democratic operatives with backing from tech employees, labor unions, and advocacy groups. Its pitch is grassroots: small donations from people who work in tech and feel their bosses are pushing AI into the world too fast and too carelessly.
The numbers are stark. The Guardrails Alliance has $5 million and hopes to raise $15 million this cycle. Leading the Future, the main pro-industry group on the other side, has raised over $140 million. OpenAI's president Greg Brockman and his wife personally donated $25 million to it. Andreessen Horowitz, one of Silicon Valley's largest investment firms, and Palantir co-founder Joe Lonsdale are also among the backers. That is not a fair fight on paper.
But the money gap is not the only thing that matters here. Leading the Future has already demonstrated the risks of being too aggressive. It spent more than $7 million in attack ads against a single New York state politician, Alex Bores, who co-wrote one of the strictest AI safety laws in the country. Rather than quietly eliminating him, the spending made him nationally famous. Six super PACs have now poured nearly $30 million into his Manhattan primary race, which takes place June 23. Anthropic, OpenAI's main competitor, has backed groups that spent over $10 million supporting Bores, turning what started as a local race into a visible proxy war between the two biggest AI companies in the world.
OpenAI has publicly tried to distance itself from the attacks. The company issued a statement saying it does not direct Leading the Future and has no visibility into its operations. But reporting by multiple outlets revealed that OpenAI's own chief political strategist helped set the PAC up, and its president is one of its largest donors. Employees pushed back internally, pressing the company's leadership at a tense meeting in May to explain the connection.
This is not just drama inside the tech world. The outcome of this political battle will shape rules that affect every business that touches AI. The central fight is between two visions: one side wants Congress to set a single national standard, which would likely be lighter and industry-friendly; the other side wants states to be able to pass their own laws right now, because Congress has not moved. In December 2025, the Trump administration signed an executive order directing federal agencies to push back against state AI laws it considers too strict. Colorado, Texas, and Illinois all have AI laws taking effect in 2026 anyway.
For a business operator, the practical question is: what rules will you eventually have to follow? The answer depends heavily on which side wins this fight. If the industry wins and a single federal standard emerges, compliance will likely be simpler but weaker. If states keep their own rules, a patchwork of requirements across different jurisdictions will be the reality, and that patchwork is already forming. Either way, the window to operate without clear AI rules is closing faster than most people outside the tech sector realise.