Anthropic donates $20M to Public First, a super PAC pushing for AI guardrails and transparency in opposition to OpenAI-backed PACs, ahead of the US midterms
Anthropic and OpenAI now have their own well-funded political groups that will square off in the midterm elections over artificial intelligence safety and regulation.
New York TimesTheodore Schleifer
Context & Ripple Effects
The donation formalizes a political vehicle for Anthropic’s AI-policy agenda at a moment when Anthropic and OpenAI are backing rival groups. The immediate significance is not merely spending, but the transfer of competition between leading AI developers into electoral advocacy.
Related coverage shows Public First later began advertising for AI regulation in New Jersey and that the two sides’ organizations became a bitter midterm influence fight, especially in Democratic primaries. That arc makes this initial funding a consequential step in building durable political capacity around competing AI-policy positions.
First-order effects
Public First gains substantial resources to advocate for AI guardrails and transparency ahead of the midterms, giving Anthropic an organized channel for its policy message.
OpenAI-backed PACs face a better-funded, explicitly opposing political organization, sharpening the contest over how AI safety and regulation are presented to voters and candidates.
Second-order effects
Candidates and party organizations in contested races may encounter more AI-policy advertising and advocacy, as later reflected in Public First’s New Jersey ad campaign.
Rival AI companies have stronger incentives to develop distinct public-policy coalitions rather than treating regulation as a shared industry agenda.
Third-order effects
If this model persists, AI governance could increasingly be shaped through campaign infrastructure and partisan coalition-building, alongside conventional lobbying and technical rulemaking.
The resulting alignment of AI firms with competing political networks may make regulatory compromise harder, while raising the value of labs that can credibly operate within changing government expectations.
The trend: AI developers are turning safety and regulatory differences into organized political competition, making electoral influence a more central layer of AI governance.
The campaign-finance details: Anthropic is voluntarily disclosing its $20 million donation to a dark-money 501(c)4 group, Public First. Public First, in turn, makes donations to its allied Democratic and Republican super PACs. Public First is going up with ads supporting
NEWS: Anthropic is putting $20 million into a super PAC operation that is meant as a direct counter to the OpenAI super PAC operation. The midterms are the new battleground between these two rival AI labs. https://www.nytimes.com/...
Anthropic is sending $20 million to Public First Action, a new group which is launching a pro-Sen. Blackburn ad on KOSA today Defending Our Values PAC is also launching a today backing Sen. Ricketts on keeping advanced AI chips out of hands of adversaries
And more! Anthropic just announced it donated $20m to Public First Action, the group which has (directly and indirectly) funded both these ads. [image]
🚨The Public First AI safety advocacy groups/PACs have launched their first ads: - One supporting Marsha Blackburn in Tennessee (praises her work on KOSA and TRUMP America AI Act) - One backing Sen. Pete Rickets (a big chip export control supporter) in Nebraska — doesn't talk
Anthropic's $20M backing pro-regulation candidates R&D World vs OpenAI's partners spending $125M to kill state AI laws. The two leading AI labs are literally funding opposing political campaigns to control how they get regulated. The companies building superintelligence are now
It's the battle of the AI PACs. Anthropic is giving $20M to Public First Action, a group backing candidates who want MORE regulation on AI and chips. Public First Action is making their first six-figure ad buys today - backing safe bets of Marsha Blackburn and Pete Ricketts.
Anthropic's announcement of a $20 million political spend to regulate AI has a strong taste of the smugness too common in the AI x-risk community. It suggests that anyone who disagrees with Anthropic's policy positions simply doesn't understand what is at stake. [image]
Anthropic has a habit of tripping over itself with the Trump administration that reminds me of Chris Moltisanti's heroin habit, where I as the sympathetic viewer am like, “my god, Chris, you were doing so well, you were on the right track, and then you blew it all up.” [image]
.@AnthropicAI is donating $20 million to Public First Action, a new bipartisan 501(c)(4) working to get AI policy right. Public First Action will work with Republicans, Democrats, and Independents who share the same policy priorities: ✅ Insisting on AI model transparency
Anthropic's $20M to Public First Action backs candidates who want model transparency, bio/cyber safeguards, chip export controls, and balanced federal rules (no blanket state preemption). Do you think this accelerates meaningful AI policy in '26, or just heats up the PAC battles?
AI is being adopted faster than any technology in history. The window to get policy right is closing. Today we're contributing $20m to Public First Action, a new bipartisan org that will mobilize people and politicians who understand what's at stake. https://www.anthropic.com/...
You couldn't even bring yourselves to officially endorse SB 1047. I hope this money can go towards serious regulation but I'm worried it will drown out calls for a global pause with milquetoast incident reporting mandates and we'll still all end up dead.
What started as an AI safety / EA-driven PAC effort to help Alex Bores in NY is apparently now becoming a vehicle for Anthropic's squabble against NVIDIA over export controls. And the PAC's first recipients are Republicans. The bedfellows are indeed strange.
A pretty interesting release from Anthropic as they raise money to make new inroads into AI. “Recent polling finds that 69% of Americans think the government is ‘not doing enough to regulate the use of AI.’ We agree.” https://www.anthropic.com/...
Pretty big contrast with OAI here, not just on substance of the respective PACs but on transparency re: the PAC being explicitly supported by the company (Ant) vs. “just” being funded by a cofounder and also totally coincidentally being closely tied to its head of policy (OAI)
One thing about the dueling AI super PACs is how BOTH sides are affected by what happened with crypto in 2024. The AI industry accelerationists explicitly want to replicate what happened with Fairshake. Some of the operatives and donors are even the same. Meanwhile, the AI
It's kind of funny to imagine all these powerful rich and smart (at least in some ways) tech people freaking out this way. Like if you really think this can you not unplug it?