OpenAI President Greg Brockman really didn’t want New York congressional candidate Alex Bores, who supported states’ right to enact AI safety regulations, to win his June primary race. But his top-donating employees felt differently — and gave hundreds of thousands of dollars to Bores and groups backing him.
OpenAI workers not only revealed a schism within one of San Francisco’s frontier AI labs, as previously reported, they also aligned with the politics of their direct competitor Anthropic, where workers and executives are walking in lockstep.
Anthropic CEO and co-founder Dario Amodei and his employees are, unlike OpenAI, on the same side politically: Both gave to the same pro-Bores and pro-federal AI safety regulation committees, according to federal filings from January 2025 to June 2026.
Anthropic and OpenAI employees also both gave hundreds of thousands in support of San Francisco congressional candidate Scott Wiener, whose AI labs regulation bill (SB 1047) was publicly opposed by OpenAI bosses.
Anthropic’s top exec and its staff are giving to pro-regulation groups. At OpenAI, the giving is split.
Contributions to pro-regulation and tech-aligned political committees
ANTHROPIC
$1M
Dario Amodei
Anthropic CEO
Anthropic
employees
Public First
AI safety PAC
$2.2M
$4M
Jobs and
Democracy PAC
$15M
$254K
Alex Bores
for NY
pro-regulation
candidate
$197K
Scott Wiener
for Congress
pro-regulation
candidate
OPENAI
$86K
OpenAI
employees
$86K
$5K
$12.5M
Leading the
Future
tech-aligned PAC
Greg Brockman
OpenAI President
$12.5M
$6M
(against)
$20M
Think Big
tech-aligned PAC
MAGA Inc.
Graphic by Kelly Waldron. Source: Federal Election Commission.
A fabled rivalry has defined both labs. Anthropic was founded in January 2021 after seven former OpenAI employees, including Amodei and his sibling Daniela, left the company over concerns that AI development had become heedless and irresponsible under founder and CEO Sam Altman.
While Altman also gave politically, it was in far smaller amounts: He gave $7,000 to Democratic U.S. Senator Mark Warner, who introduced the pro-regulation Secure A.I. Development Act this year. OpenAI employees backed that bet with another $18,000.
The political giving has drawn fault lines in the battle over AI safety within the country’s emergent AI donor class, many of whom are poised to become ridiculously rich as OpenAI and Anthropic aim for trillion-dollar valuations.
U.S. leaders have publicly mulled everything from AI kill switches to new laws requiring disclosure of security breaches since OpenAI, Anthropic and Meta — in their frenzied battle for AI dominance — caused public alarm when AI agents, on multiple occasions in the last month, escaped their testing sandboxes and hacked the servers of other companies and websites.
In May, Anthropic co-founder Amodei wrote a $1 million check to Public First, a bipartisan political action committee backing candidates who support AI guardrails. This year, the committee mainly backed Bores, an ex-Palantir worker who championed a state law imposing transparency, safety, and reporting requirements on frontier labs.
Five of Amodei’s workers — research scientist Adam Jermyn and technical staff Peter Lofgren, Nicholas Joseph, Jan Leike, and Shauna Kravec — opened their wallets for Public First as well.
In fact, they outspent Amodei with a collective $2.2 million to Public First, which subsequently spent on a committee that backed Bores with $13 million. Separately, 45 Anthropic employees gave a collective $254,000 to Bores directly. Daniel Ziegler, a machine learning researcher at Anthropic, gave $790,000 to another pro-Bores committee, Dream NYC.
Meanwhile, OpenAI co-founder and president Brockman, and his wife, Anna, gave a combined $50 million to two super PAC networks, Leading the Future, which supports AI-friendly candidates from both political parties, and MAGA, Inc., a pro-Trump super PAC.
Leading the Future, a committee funded by anti-regulation tech figures, opposed Bores’ congressional campaign with $6 million in independent spending through another pro-AI, anti-regulation committee called Think Big.
But none of Brockman’s employees matched his giving. Instead — like Amodei and his employees at Anthropic — they gave to pro-regulation PACs. Two OpenAI workers gave a combined $205,000 to Guardrails Alliance, which spent nearly $300,000 in support of Bores—a committee directly opposing Brockman’s. Juan Felipe Ceron Uribe, an OpenAI technical staff member, also gave $200,000 to Dream NYC to support Bores.
On top of that, 16 OpenAI employees collectively gave $86,000 in direct donations to Bores. Seventeen of them also gave a collective $93,000 directly to Wiener, and thirty-four Anthropic employees collectively gave $200,000 to support Wiener.
“The employees at OpenAI are not blindly following what their employers are doing,” said Zhao Li, an assistant professor at New York University who studies corporate political activities in nascent industries like AI.
These are people “who are smart enough to work at two of the most consequential firms on our planet, so they command quite a bit of power in the labor market; they have talent that’s hard to replace; and they might be on the same side on issues such as AI safety,” Li said.
“We’re looking at a subset of workers that have strong political opinions, have the information to act on them and the money to give and presumably have job security.”
A spokesperson for OpenAI declined to comment on the data, deferring to a company statement on workers’ political giving: “Our employees are free to participate in the political process in their personal capacities, including by donating or providing advice to candidates, campaigns, and political organizations. When they do that, they speak for themselves and not OpenAI.”
Other frontier labs have seen their workers enter the political ring. Employees at the Google DeepMind lab, whose core teams are in the Bay Area, gave at least $320,000 to federal races between this and last year. Most of that money, $250,000, came from one person — DeepMind software engineer Kate Woolverton — and went to the pro-regulation Public First. About $19,000 from four DeepMind employees also went to Wiener’s congressional campaign in San Francisco between this and last year.
Alex Turner — a former DeepMind AI safety research scientist who went viral in July when he publicly resigned over the company’s AI contract with the Pentagon — was behind at least half of those pro-Wiener donations.
Turner said he partly opened his wallet over concerns about the dearth of U.S. laws requiring frontier labs to notify the federal government of cybersecurity incidents.
“There’s no law that says companies have to tell them about these events,” Turner told Mission Local. “You could have these extremely powerful entities growing in the United States that are out of control of their creators and regularly compromising internal systems. Congress should know about it and I think the American people should know about it too.”
Turner said there remains a gap between the average voter and AI as a public safety priority.
“People are reasonably preoccupied with putting food on the table and making car payments,” Turner said. “When they hear about cybersecurity incidents involving companies named Hugging Face — it sounds kind of silly. Sadly, this is quite serious. OpenAI’s negligence today might endanger everyone tomorrow.”
