I’ve been watching a quiet shift inside the tech world that rarely makes it into the shiny product demos: everyday engineers, designers, and researchers are starting to push back—openly—against how their companies build and ship artificial intelligence. Not in the abstract, either. They’re asking for something bluntly practical: **develop AI responsibly, and stop using politics to dodge accountability**. Out of that mood comes a new player that wants to turn workplace frustration into election-season leverage.
The vehicle is the Guardrails Alliance, a newly launched super PAC focused on supporting AI legislation. According to reporting by The New York Times, Democratic political operatives Shaunna Thomas and Leah Hunt-Hendrix introduced the group with backing from tech workers, labor unions, and other allied organizations. If you’ve ever wondered what happens when internal Slack debates about “ethics” collide with the hard realities of Washington, this is it.
From employee unease to an explicit political project
Thomas described the mission in stark terms, arguing that **people still have the power to stop an “authoritarian” capture of politics and technology** linked to Trump-era forces and a deregulatory tech agenda. Strip away the rhetorical heat and you’re left with a clear strategic bet: the AI boom has created not only new fortunes, but also a new class of workers who feel exposed—professionally and morally—by what their tools can do in the real world.
And let’s be honest: for a long time, the tech industry has acted as if “responsible AI” is best handled through internal principles and polished blog posts. But what if the decisions that matter most—liability rules, disclosure requirements, safety standards, enforcement budgets—are being decided by lawmakers who mostly hear from the best-funded voices in the room?
That’s the opening Guardrails Alliance is trying to exploit: a sense that **the public conversation is being purchased**, and that even many people inside the companies building AI don’t like what they’re seeing.
Small-donor “populism” versus deep-pocketed influence
Guardrails is positioning itself as a small-donor, grassroots-style political effort—an attempt to build a “political home” for workers and voters who want tougher rules for AI. Right now, the PAC reportedly has around $5 million and says it aims to raise $15 million during this election cycle.
In modern super PAC terms, that’s real money but not empire-building money. The contrast becomes obvious when you put it next to Leading the Future, a competing group that reportedly has more than $100 million. That war chest is said to include money from major tech figures, including OpenAI president Greg Brockman.
If you’re doing the cynical math like I am, you’re probably asking: can $15 million matter against $100 million-plus? Thomas’s response is essentially that Guardrails isn’t trying to win a spending contest. The pitch is different: **organize the people who feel pressured, ignored, or outright steamrolled by an anti-regulatory AI industry**, and then use targeted spending to make that constituency visible to politicians.
A New York race becomes a test case
The first concrete move is in New York, where Guardrails plans to buy ads supporting Alex Bores, a congressional candidate facing a primary next week. Bores is notable here because he became the first target of Leading the Future—an early signal that AI policy is no longer a niche issue buried in committee hearings. It’s becoming campaign ammunition.
On the same day Guardrails was launched, Bores shared an advertisement featuring the parents of Adam Raine, a teenager whose death by suicide followed months of extended conversations with ChatGPT. It’s the kind of gut-level story that drags AI harms out of the realm of hypothetical risk and into the uncomfortable space of accountability: what guardrails exist, who sets them, and what happens when systems designed for engagement collide with vulnerable users?
Bores is also being supported by another pro-legislation super PAC, Public First Action, which is backed by Anthropic. So the landscape is messy in a very 2026 way: competing political groups, overlapping alliances, and AI labs appearing—directly or indirectly—in the funding ecosystem around “responsible” regulation.
OpenAI, donations, and the credibility problem inside tech
One detail that sticks with me is the reported internal skepticism among OpenAI employees. Even though OpenAI has attempted to distance itself from Brockman’s donations, many workers apparently aren’t convinced. Some have voiced concern on social media about the attacks Leading the Future has directed at Bores.
This is the credibility trap for AI companies right now: they want to be seen as builders of world-changing tools and also as trustworthy stewards. But when prominent leaders’ money shows up in aggressive political operations—especially ones framed as anti-regulatory—employees notice. And the public notices the employees noticing.
It’s not simply a question of partisan politics. It’s about whether “AI safety” is a sincere commitment or a branding layer that peels off the moment regulation threatens margins.
The broader worker push: ICE contracts and Pentagon pressure
The Guardrails story is also part of a wider pattern: tech workers increasingly organizing around where their companies’ tools end up. This year, employees have mobilized to demand that leadership end contracts with U.S. Immigration and Customs Enforcement (ICE). In parallel, workers have urged the Pentagon to remove Anthropic’s designation as a supply-chain risk.
Critics argue that the Pentagon label was imposed without due process and functioned as retaliation—specifically because Anthropic has placed limits on using its technology for mass surveillance and autonomous warfare. Whether or not every detail survives scrutiny, the underlying tension is unmistakable: **AI is now a national-security object, a corporate product, and a moral hazard—often at the same time**.
If you’re an employee in the middle, what do you do? Quit? Whisper concerns internally and hope someone listens? Or try to build political power that makes it harder for executives and agencies to make consequential decisions behind closed doors?
What Guardrails is really trying to build
Thomas framed Guardrails not as an attempt to outspend opponents, but as a structure—an actual “home”—for people who think the AI industry’s deregulatory wing is trying to manipulate elections. That’s a sweeping claim, but it reflects a very real fear: that massive, targeted political spending can define what “AI policy” even means in the public imagination.
Guardrails Alliance is, at its core, a bet that the coalition of concerned tech workers, unions, and allied groups can become something more than a protest. The question for you—and for me—is whether this becomes a lasting counterweight, or just another brief flare in a politics-and-tech cycle that chews through movements as quickly as it ships new models.
TechCrunch has reached out to Guardrails Alliance for comment.
Comments