AI Safety Advocates and Industry Giants Fuel Midterm Campaigns with Millions

Groups on both sides of the artificial intelligence safety debate are injecting substantial funds into the upcoming midterm elections, with a significant portion directed towards candidates supporting increased regulation of the technology. Reports indicate at least $48 million has been allocated to political campaigns by these entities.

The Guardrails Alliance, an AI safety advocacy group supported by tech workers, has reportedly channeled $11.7 million towards candidates who favor stricter regulations on AI development. This effort is part of a broader trend where individuals and organizations concerned with AI's potential risks are making significant political contributions.

Juan Felipe Cerón Uribe, associated with OpenAI and known for his focus on mitigating catastrophic AI risks, has personally contributed approximately $400,000 to the political action group Dream NYC. This PAC, in turn, previously allocated around $2.5 million to Alex Bores, a New York State Assemblyman and proponent of AI safety, for his unsuccessful campaign.

Dario Amodei, CEO of Anthropic, a company that has raised concerns about AI's trajectory, has committed $1 million to Public First, a political group backing candidates who advocate for slowing down or regulating AI advancements. Public First Action has distributed funds to various PACs supporting a range of candidates, including $1.6 million to North Carolina Representative Valerie Foushee and nearly $1.5 million to Texas congressional candidate Carlos De La Cruz.

Furthermore, AI safety "doomers" have donated at least $41,000 to Peggy Flanagan, a Senate candidate in Minnesota endorsed by Sen. Bernie Sanders, who has characterized AI as an "existential threat" to humanity.

Conversely, groups advocating for less regulation and promoting technological innovation are also making substantial financial commitments. Venture capitalist Marc Andreessen and OpenAI President Greg Brockman have collectively invested $37.5 million into the super PAC Leading the Future. This PAC has amassed a total of $149 million and is supporting candidates such as Melissa Bean, a pro-business Democrat in Illinois.

These financial flows are occurring against the backdrop of an intensifying debate over AI safety. Leading technology companies, including Anthropic and OpenAI, have publicly called for development slowdowns while simultaneously pushing the boundaries of their own research. Concerns about AI's potential negative impacts on humanity, coupled with opposition to the vast data centers required for AI, are growing louder across the political spectrum.

Aryan Bhatt, who works at Redwood Research, an organization with noted connections to Anthropic, has contributed to congressional candidates like House Republican Bill Huizenga, an advocate for AI guardrails centered on national security. Bhatt's group, Redwood Research, co-authored a report detailing a simulated hack of Hugging Face by rogue OpenAI agents.

Guardrails Alliance has also directed funds against candidates perceived as promoting unchecked tech growth, including a $412,500 contribution against Republican Aaron Flint in a Montana House race.

Donors supporting AI safety initiatives state their motivation is not partisan but rather a response to the significant investments made by tech giants in electing pro-innovation candidates. Leading the Future has indicated its goal is to build a broad coalition, recognizing that 60 Senate votes would be necessary to pass stringent AI regulation.

The polarizing nature of AI regulation has led to some donor hesitation. Greg Brockman reportedly rescinded a planned $25 million donation from Leading the Future following internal dissent at OpenAI.