Anthropic, the company behind Claude AI, says it’s deeply committed to AI safety and responsible development. Anthropic also just created a political action committee to fund candidates in the upcoming midterms. If that sounds contradictory, you’re paying attention.
The newly formed AnthroPAC will make bipartisan contributions to both current lawmakers and rising political candidates during the midterm elections. The PAC will be funded exclusively through voluntary employee donations, which is a detail worth examining. When your employees are writing checks to influence policy that affects your business, the line between grassroots activism and corporate lobbying gets pretty blurry.
Following the Money
This isn’t Anthropic’s first foray into political spending. Back in February, the company donated $20 million to Public First Action, a group launched specifically to support AI safeguard efforts. That’s a substantial sum, especially for a company that positions itself as the thoughtful alternative in the AI space.
So what’s actually happening here? On one hand, you could argue that Anthropic is putting its money where its mouth is. If you genuinely believe AI needs guardrails, then influencing the people who write those guardrails makes sense. On the other hand, every tech company that’s ever lobbied Congress has claimed they were doing it for the greater good.
The Bipartisan Approach
AnthroPAC plans to contribute to candidates from both parties, which is standard operating procedure for corporate PACs. You don’t want to bet on just one horse when either party could end up controlling the committees that regulate your industry. This bipartisan strategy is practical, but it also means the PAC isn’t really about ideology or specific policy positions. It’s about access.
For those of us trying to understand AI agents and their implications, this development matters because it shows how quickly the AI industry is maturing politically. We’ve gone from “move fast and break things” to “move fast and make sure the people writing the rules are people we’ve donated to.”
What This Means for AI Regulation
The timing is significant. AI regulation is still being figured out at the federal level, and whoever gets in early has a better chance of shaping those rules. Anthropic clearly understands this, which is why they’re not waiting around for someone else to define what “responsible AI” means in legal terms.
But here’s the tension: can a company genuinely advocate for strict AI safety measures when it’s simultaneously funding politicians who might water down those same measures? The $20 million to Public First Action suggests Anthropic wants real safeguards. The creation of a PAC to fund individual candidates suggests they want to make sure those safeguards don’t hurt their business model.
Maybe both things can be true. Maybe Anthropic really does want smart regulation that protects people without stifling development. Or maybe this is just what happens when an AI company grows up and realizes it needs to play the same political games as every other tech giant.
The Bigger Picture
For non-technical folks trying to make sense of the AI space, this is an important signal. When AI companies start forming PACs, it means they’re no longer scrappy startups. They’re established players with established interests to protect. That doesn’t make them evil, but it does mean we should probably stop taking their safety rhetoric at face value.
The question isn’t whether Anthropic should be involved in politics. The question is whether their political involvement will actually lead to better AI safety outcomes, or just to regulations that happen to benefit Anthropic. Based on how corporate PACs have worked in every other industry, I’m not holding my breath for the former.
🕒 Published: