AI Claims Political Neutrality—But ‘Election Safeguards’ Partners Are Flush With Left-Leaning Cash

The tech firm Anthropic asserts it wants to prevent political bias on its AI platform Claude, but significant funding from left-leaning organizations has gone to three groups providing feedback on the election project.

One of the groups is a center-right think tank funded by some left-leaning donors, while another partner in Anthropic’s “election safeguard” program has numerous advisory council members with significant activism experience. A think tank at Vanderbilt University has also received grants from center-right organizations.

“We train Claude to treat different political viewpoints equally and work with experts and institutions whose affiliations and funding span the political spectrum,” an Anthropic spokesperson told the Daily Signal. “We test extensively for bias before every model launch and both our internal evaluations and independent third-party assessments consistently show that Claude provides balanced responses on political topics.”

Donors and Advisers on the Left

Donors to one of the three groups, the Collective Intelligence Project, include the Omidyar Network, a nonprofit established by eBay founder Pierre Omidyar that is part of the larger Omidyar group of nonprofits.

Omidyar also established the Democracy Fund, which has contributed to left-of-center get-out-the-vote organizations. His Democracy Fund Voice contributes to left-of-center political organizations, including the Sixteen Thirty Fund. The Omidyar Network announced in 2024 that it bought about 50,000 shares of Anthropic.

The Ford Foundation is a longstanding institutional funder of left-of-center causes and has contributed to CIP, according to the CIP 2024 annual report.

In March 2024, the Ford Foundation bought $5 million worth of Anthropic shares, the New York Post reported.

Further, CIP’s Advisory Council includes Eli Pariser, a former executive director of MoveOn.org, a left-wing political group; Kinney Zalesne, the deputy national finance chair of the Democratic National Committee; Katherine Maher, the president of National Public Radio; and Harvard professor Danielle Allen, who ran for governor of Massachusetts in the 2022 cycle as a Democrat.

A spokesperson for the Collective Intelligence Project did not respond to inquiries for this story.

Balance on the Right

The Foundation for American Innovation website notes it has been funded by a combination of grant makers on both the right and the left.

This includes a $1.7 million grant in 2024 from the Packard Foundation, which largely funds environmental and birth control projects.

Two grants totaling $976,000 came from Coefficient Giving, an organization backed by major Democratic donor Dustin Moskovitz. Moskovitz, who co-founded Facebook, has a reported net worth of $15 billion. He gave $38 million to the Future Forward super PAC supporting Kamala Harris’ presidential campaign in 2024, the Financial Times reported.

The William and Flora Hewlett Foundation reported making one contribution of $225,000 and another for $125,000 in 2024. The Hewlett Foundation has supported Planned Parenthood, the Sierra Club, and other left-leaning organizations.

Still, Inc. magazine has characterized the think tank as being on the right. It is a member of the conservative State Policy Network, which is a coalition of more than 200 conservative and libertarian think tanks across the United States.

Notably, conservative-leaning donors such as the Lynde and Harry Bradley Foundation and the Sarah Scaife Foundation each made contributions of more than $300,000 per year in 2024 and 2023, according to the Capital Research Center, which monitors nonprofits.

“FAI draws support across the philanthropic map because we are the leading national tech and AI policy think tank, not just on the political right,” FAI President Zach Graves told the Daily Signal in an email statement.

He noted no single funder provides more than 10% of the group’s revenues, and the support base is “remarkably diverse.”

“Importantly, AI policy is an area with intense intra-party disagreement on both sides, with policy debates that aren’t neatly polarized across traditional left vs. right lines,” Graves added. “The work we produce helps a range of funders shape their thinking about AI—not the other way around.”

Vanderbilt Think Tank

The third partner organization is the Future of Free Speech, based at Vanderbilt University in Nashville, Tennessee.

“We engaged with Anthropic because of our ongoing research evaluating the free speech culture of AI chatbots — specifically, whether AI company content moderation policies align with freedom of expression principles and whether chatbots are willing to generate content on controversial subjects,” Jordi Calvet-Bademunt, senior research fellow at the Future of Free Speech, told the Daily Signal in a statement.

He said a 2024 report found that chatbots were more likely to refuse prompts that would have generated a right-of-center response than those that would have generated a left-of-center response. But a 2025 follow-up study found an overall reduction in refusals across controversial topics.

“For the past two years, we have produced two reports that measure the willingness of prominent chatbots to respond to controversial prompts,” he said. “The topics we tested in these reports span the political spectrum, including transgender athletes in women’s sports, the criminalization of abortion, the COVID-19 Chinese lab leak theory, the Israel-Palestine conflict, systemic racism in the U.S., and the role of European colonialism in current global inequalities.”

The organization received a $35,144 grant in 2024 from the Foundation for Individual Rights and Expression, or FIRE, a libertarian-leaning free speech group, according to Cause IQ, which tracks nonprofits. It also received a $25,000 grant from the free-market Atlas Network in 2024.