U.S. Campaign Finance Disclosures Show OpenAI Spending Despite Election Ad Limits

Two candidates said explicitly in filings that they had used the subscription for advertising, despite an OpenAI policy that prohibits candidates from using its tools to generate ads. Another candidate disclosed using AI to draft and personalize political messages or create synthetic media, though the filing did not specify which software was used. About 30 political action committees and parties also reported payments to OpenAI, with the Republican National Committee the company’s largest political spender at roughly $9,700.

Political consultants say such disclosures understate how many candidates rely on ChatGPT and other AI tools. Eric Wilson, a Republican digital strategist who has advised campaigns on AI, said the tools are mostly used for writing emails, ad copy and scripts, but that no one would go around saying I am using AI. That lack of transparency reflects a paradox: generative AI has become ubiquitous, and candidates could be at a disadvantage if they avoid it. Studies have found voters are less likely to trust messages they know were generated by AI, which makes campaigns reluctant to disclose its use. Political consultants expect that as Election Day approaches, more campaigns will outsource AI-generated ads and materials to super PACs, much as they already do with negative ads.

Katie Harbath, CEO of the tech policy consulting firm Anchor Change and a former Meta executive, said there are many parallels between AI-generated content and negative campaign ads. Voters say they find such ads distasteful, she said, but politicians keep using them because they work. Campaigns typically give riskier material to outside groups, she said. Harbath also described a growing tension on the left over the use of AI, with Democrats saying that if Republicans are using it, they should too. Not using AI in voter-facing materials, she warned, could become a serious disadvantage.

AI companies have developed policies to prevent targeted disinformation, but The Post’s analysis found OpenAI unevenly enforces its restrictions, making it possible for campaigns to circumvent election rules. In late July and early August, the newspaper asked ChatGPT to generate fundraising text messages targeting mothers on behalf of a female veteran running for office; the chatbot produced multiple tailored texts in an apparent violation of policy. When asked with the same prompt later, the chatbot declined, saying it could not draft political persuasion or fundraising messages targeting a demographic group such as moms. The chatbot also inconsistently enforced rules that prohibit campaigns from using its tools to write emails to voters, at times complying with a request to write a fundraising email for a specific candidate and at other times denying the same request on the same day.

Wilson said OpenAI should get more feedback from political consultants and campaigns on its policies, since the rules can sometimes seem arbitrary or contradictory. He said it makes no sense that a campaign can use ChatGPT to develop its policy on early-childhood education but not to create a social media post promoting that policy. A survey by the American Association of Political Consultants earlier this year found that 57 percent of consultants reported using AI for their work daily, up from 34 percent a year earlier.

The Post noted that election-related deepfakes have so far been quickly debunked or gained little traction in U.S. elections. More than 30 states have enacted a patchwork of laws limiting how politicians can use deepfakes in campaigns, but states often have limited resources to enforce the laws, and some have been challenged as unconstitutional.