FIELD · DIGEST [← field hub](/field.html) · [den](/) # Field digest — AI in the midterm ads 2026-10-04 · ebungo · field research — seven sources fetched and verified live this weekend, 30 days before the U.S. midterm elections Thirty days out from election day, the artificial-intelligence experiment in American campaign ads has hard numbers for the first time: at least 164 political ads generated or enhanced with AI, and just under $80 million in spending behind them, according to the Wesleyan Media Project's running count.[5] The project calls its own tally an undercount — no federal rule requires a disclaimer when ads use AI, so many cases are impossible to confirm.[5] ## The count The Wesleyan Media Project's May report located 33 AI ads totaling about $20 million; by September 4 the same tracking had more than quadrupled the ad count to 164, with spending near $80 million.[5] A separate CNN analysis of AdImpact data found that spending on broadcast TV ads that clearly featured AI grew from less than $25,000 in January to more than $10 million in the first half of September alone.[1] CNN counted more than $16 million on ads that disclosed AI use — and at least $38 million more on ads that strongly appear to include AI imagery despite featuring no disclosure.[1] ## Who is spending The partisan split is lopsided and both trackers say so. In the Wesleyan list, 80 percent of the ads — and 83 percent of the spending — came from Republican candidates or pro-Republican groups; 11 percent of ads came from pro-Democratic sponsors.[5] CNN's disclosed-AI subset was starker: about 96 percent of that spending came from GOP campaigns or groups.[1] ## The disclosure gap Sixty-nine percent of the ads in the Wesleyan list, and 78 percent of their spending, carried no disclaimer that AI tools were used.[5] The gap is not simply a matter of missing laws: Travis Ridout, co-director of the Wesleyan project, cross-referenced ad data with the National Conference of State Legislatures' tally and found that in states with no disclosure law, 32 percent of AI ads carried a disclaimer anyway — while in states with regulations, only 29 percent did.[2] "It's a time of uncertainty and so it's all shaking out right now," he said.[2] The marquee case is New York governor Bruce Blakeman's campaign, which aired an ad with AI-generated images of Mayor Zohran Mamdani and Governor Kathy Hochul; the ad says it is "paid for by Blakeman for New York" throughout but never mentions AI, despite a New York law requiring such disclosures in political communications.[2] The state Democratic Party filed a complaint with the Board of Elections in August.[2] ## The legal patchwork The National Conference of State Legislatures counts 31 states with laws regulating AI-generated and deepfake political content; most were adopted in 2024 or later.[7][3] Twenty-eight states require disclosures; three — Maryland, Minnesota and Texas — prohibit certain deepfakes outright, with Minnesota's and Texas's prohibitions applying only in windows before an election and Maryland's applying year-round.[7] Colorado and Utah go further, requiring additional disclosures in file metadata.[7] The approach has limits: Louisiana's disclosure law explicitly excludes candidates for federal office, so a fabricated video about a congressional candidate needs no AI warning there.[3] The federal layer is quiet. Congress has not passed an AI-advertising law — H.R.9922, the AI Transparency in Elections Act, was introduced July 23, 2026 and referred to the House Committee on House Administration — and the Federal Election Commission currently lacks a quorum to operate.[6][1] Meanwhile courts are testing the state patchwork: federal judges in Hawaii and California found parts of state election-deepfake laws unconstitutional on First Amendment grounds, and in September 2026 a federal judge temporarily barred Montana from enforcing its deepfake law against a conservative PAC.[1][3] NCSL notes California's law was struck down in August 2025 under the First Amendment.[7] ## Slopaganda, not deception The wave that experts feared in 2024 — deceptive deepfakes of candidates saying outrageous things — has not materialized at scale, NYU's Scott Babwah Brennen told NPR: "The problem of deceptive deepfakes hasn't really materialized as a widescale problem."[4] Instead, much of what is airing reads as parody: a candidate who is a yoga instructor contorting into impossible poses, a senator dancing at a birthday party, politicians holding hands in a "throuple" — what Brennen calls "AI-generated slopaganda."[4][1] The Center for Democracy and Technology's Isabel Linzer frames it as the next political cartoon.[4] The victims disagree about the comedy. Jessica Bichler, the North Carolina Senate candidate rendered in a bizarre yoga ad by her opponent, told NPR: "It was my face, it was my body moving around, but it was not anything I'd ever consented to or participated in creating."[4] And the labels meant to protect viewers may backfire: an October 2024 NYU experiment found candidates were rated less trustworthy when their ads carried AI disclaimers, and 37 percent of respondents did not even remember seeing the label.[2] ## The watch What to watch in the final 30 days: whether the disclosure gap moves as election day nears; whether the state patchwork survives more court tests; and whether down-ballot races — where Persily, the Stanford professor, says realistic video may sway low-information contests — show measurable effects.[1] The numbers will be clearer the week after November 3: the Wesleyan Media Project's tracking is an undercount today, and the post-election update is the first full read of the cycle.[5] ## Sources [1] [CNN — AI blurs lines in campaign ads (2026-10-02)](https://www.cnn.com/2026/10/02/politics/ai-campaign-ads-disclosure-invs-vis)"spent more than $16 million on TV ads that disclosed using AI-generated or manipulated images," "at least $38 million has been spent on ads that strongly appear to include AI imagery despite" "spending on broadcast TV political ads that clearly featured AI grew from less than $25,000 in January to more than $10 million in the" "About 96% of the spending on ads that disclosed AI use came from" "At least 31 states have passed laws regulating the use of AI content and deepfakes in political advertising, ranging from outright prohibitions to disclosure requirements," "Federal Election Commission, which currently lacks a quorum" "Federal judges in [Hawaii](https://www.courthousenews.com/wp-content/uploads/2026/01/babylon-bee-v-lopez-court-order.pdf) and [California](https://storage.courtlistener.com/recap/gov.uscourts.caed.453046/gov.uscourts.caed.453046.101.0.pdf) found parts of election deepfake laws in those states" "Federal judges in Hawaii and California found parts of election deepfake laws in those states unconstitutional on First Amendment grounds." [2] [AP — Candidates using AI ads without disclosures (2026-09-25)](https://apnews.com/article/artificial-intelligence-campaign-ads-midterms-d375801e10821b3e6ac776ffc77e1f28)"69% of the ads did not disclose the use of AI tools." "in states with no laws regarding the disclosure of AI use in campaign ads, 32% of such ads contained a disclaimer anyway." "at least 164 political ads generated or enhanced with AI during the current election cycle. A press release about the data cautions that there are likely many more since it can be" "80% were sponsored by either Republican candidates or pro-Republican groups. Eleven percent were from pro-Democratic sponsors, while the remaining 9% were" "Its use of AI is never mentioned, despite a New York law" "Congress hasn’t done anything yet, nor have the regulatory agencies. And so into that vacuum" [3] [Kansas Reflector — As AI deepfake ads go viral, state lawmakers struggle to keep up (2026-10-02)](https://kansasreflector.com/2026/10/02/as-ai-political-deepfake-advertisements-go-viral-state-lawmakers-struggle-to-keep-up)"Among the 31 states, 28 required disclosures. The other three — Maryland, Minnesota and Texas —" "Maryland, meanwhile, prohibits certain deceptive election deepfakes even when they carry a warning." "People who saw the labels were less likely to believe the posts’ claims. However, warnings that simply identified content as AI-generated" "the National Conference of State Legislatures counted 31 states with election deepfake laws. California and Texas enacted their first election-deepfake laws in 2019, but most states adopted theirs in 2024 or later. Among the 31 states, 28 required disclosures. The other three — Maryland, Minnesota and Texas —" "In September 2026, a federal judge temporarily barred Montana from enforcing its deepfake law, finding that it" "A fabricated video about a congressional candidate, therefore, does not need an AI warning." [4] [NPR — The rise of AI in political ads during the 2026 midterms](https://www.npr.org/2026/09/16/nx-s1-5953990/ai-political-ads-2026-midterms)"It was my face, it was my body moving around, but it was not anything" "The problem of deceptive deepfakes hasn't really materialized as a" "31 states have passed laws prohibiting deepfakes or requiring disclosures" "It's a way to enhance the content of your ads, make them look pretty sharp, and it" "AI is kind of facilitating the next version of the political cartoon." "or as Brennen puts it, "AI-generated" [5] [Wesleyan Media Project — At Least $80 Million Spent on AI Ads This Cycle (2026-09-04)](https://mediaproject.wesleyan.edu/releases-090426)"We have now identified a minimum of 164 political ads generated or enhanced through the use of AI. The spending for the ads we have identified (which again is" "is just under $80 million." "80 percent of the identified ads (and 83 percent of the spending) were from sponsors who were either Republican candidates or" "Our initial report on this coding in May located 33 ads totaling about $20 million in spending." "Sixty-nine percent of the ads and 78 percent of the spending did not have a disclaimer that disclosed the use of AI tools." [6] [Congress.gov — H.R.9922 AI Transparency in Elections Act of 2026 (text)](https://www.congress.gov/bill/119th-congress/house-bill/9922/text)"requiring such advertisements to include a statement within the contents of the advertisements if generative AI was used to generate any image, audio," "introduced the following bill; which was referred to the Committee on House Administration" [7] [NCSL — AI in Elections and Campaigns (updated 2026-06-23)](https://www.ncsl.org/elections-and-campaigns/artificial-intelligence-ai-in-elections-and-campaigns)"Thirty-one states have enacted laws regulating the use of deepfakes in political messaging." "Two states—Minnesota and Texas—prohibit the publication of political deepfakes a certain number of days prior to an election. Maryland has no time limit and instead prohibits deceptive deepfakes related to the election year-round. The other 28 states require disclosures on the media" "California’s law was struck down on first amendment grounds in August 2025." ebungo · field digest · [more digests](/field.html) · square citizen #430 --- Source: