Employees at Anthropic, OpenAI and Google DeepMind are pouring millions into US political campaigns and political action committees that support AI regulation, according to The San Francisco Standard. The movement is becoming an organised counterweight to more than $100m in industry spending aimed at shaping the debate, making AI safety an electoral force rather than merely a lab concern.
Watch Desk analysis
What happened
The Standard reports that rank-and-file employees, researchers and AI-safety advocates have given tens of millions of dollars to candidates and PACs over the past two years. Public First Action, funded largely by Anthropic employees, has raised more than $2.6m from employees of Anthropic, OpenAI and Google DeepMind, while the Guardrails Alliance has raised more than $700,000, including donations from seven OpenAI employees.
The political spending crosses party lines. Public First Action has announced $7m in spending supporting Republican senators Tom Cotton and Josh Hawley, alongside Democratic candidates. Other reported recipients include California state senator Scott Wiener, Jonathan Kreiss-Tomkins in Alaska and Manny Rutinel in Colorado. The movement formed around support for regulation, but donors disagree over which candidates and policies deserve backing.
Why it matters
The practical consequence is influence. Candidates who once treated AI policy as a specialist concern now face a highly motivated donor network that believes control of advanced systems is one of the defining political issues of the decade. The Standard quotes insiders who expect the movement to grow as AI companies go public and employees become wealthier.
This also raises an awkward democratic question. People working inside the companies building frontier systems may have unusually useful insight into their risks, but they are also participating in the political contest over the rules that will govern their employers. Expertise and vested interest have arrived at the same fundraising dinner.
Our read
This is a proper shift in the AI story. The argument over regulation is no longer only between executives, lobbyists and campaign groups. Employees and safety researchers are building their own political machinery, with enough money and organisation to affect races.
The Standard’s reporting establishes donations, named groups and campaign activity, not that these donors are right about the probability of catastrophic AI risk or that their preferred policies will work. The useful takeaway is simpler: AI governance is becoming an organised electoral constituency, and candidates will have to decide whether to court it, challenge it or pretend it is not there until the cheques arrive.
What to watch
- Whether AI-safety donors become a durable force beyond the current election cycle.
- Which regulation proposals attract support across party lines.
- Whether industry-backed and safety-backed spending turns AI policy into a more polarised issue.
- Whether candidates disclose clearer positions on model testing, export controls and independent evaluation.
Discussion spark: Should people working inside AI companies have a louder political voice on safety rules, or does their industry connection make that influence too conflicted to trust?
Sources and evidence
- They build frontier AI. Now they’re donating millions to rein it in – The San Francisco Standard (24 September 2026, 13:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.