> ## Content Index > Fetch the complete content index at: https://ai-in-context.ghost.io/llms.txt > Use this file to discover other available public pages before exploring further. # The Slippery End to Truth in Political Advertising - URL: https://ai-in-context.ghost.io/the-slippery-end-to-truth-in-political-advertising/ - Published: 2026-08-23T15:20:36.000Z - Updated: 2026-08-23T15:20:36.000Z - Author: Jenny Stromer-Galley As a matter of principle, I have long believed that the public should not be deceived by political campaigns. The core principle I borrow from [Jürgen Habermas](https://en.wikipedia.org/wiki/The%5FStructural%5FTransformation%5Fof%5Fthe%5FPublic%5FSphere?ref=ai-in-context.ghost.io) and [Anthony Downs](https://en.wikipedia.org/wiki/An%5FEconomic%5FTheory%5Fof%5FDemocracy?ref=ai-in-context.ghost.io) is this: the public needs quality, truthful information so that they can make rational decisions that reflect their interests. From the perspective of a healthy [representative democracy](https://www.sup.org/books/politics/models-democracy-3rd-edition?ref=ai-in-context.ghost.io) that's a necessary precondition for electoral decision-making. I still remember the scandals around a set of ads that George H.W. Bush’s campaign ran in 1988 against Michael Dukakis that portrayed him as weak on crime and[ national security](https://www.politico.com/magazine/story/2013/11/dukakis-and-the-tank-099119/?ref=ai-in-context.ghost.io) through intentionally misleading, emotion-laden, and [racially charged](https://www.history.com/articles/george-bush-willie-horton-racist-ad?ref=ai-in-context.ghost.io) messages.[\[1\]](#%5Fedn1) ![](https://storage.ghost.io/c/b3/f4/b3f40ad0-1485-4466-8b7d-5c45675dde69/content/images/2026/08/data-src-image-91869326-da56-4106-bf69-b06973338e66.png) *Source:* [*The Washington Post*](https://www.washingtonpost.com/news/monkey-cage/wp/2016/01/06/its-time-to-stop-the-endless-hype-of-the-willie-horton-ad/?ref=ai-in-context.ghost.io) Those ads today would be run-of-the-mill in our say anything, maximum outrage political landscape. Artificial intelligence makes it even easier to quickly produce deceptive ads and then throw them in the circulatory system of social media for maximum effect. Take, for example, New York’s Republican nominee for governor, Bruce Blakeman, whose campaign has released [at least 17 political ads ](https://www.newsday.com/news/region-state/blakeman-hochul-campaign-ads-a2rqxqpx?ref=ai-in-context.ghost.io)that have partially or fully AI-generated content. In one video, an AI-generated Blakeman is portrayed as a superhero who flies away like Superman. But other ads are much more problematic because they reproduce the voice and likeness of politicians from the Democratic Party. In one video, New York City mayor Zohran Mamdani’s voice and likeness is [depicted as a clerk at a grocery store ](https://x.com/NassauExec/status/2086856349670047857?ref=ai-in-context.ghost.io)demanding ID in his “government-run” grocery store. In another, the voice and likeness of current Governor Kathy Hochul is seen [walking in a city park ](https://x.com/NassauExec/status/2081733845812908447?ref=ai-in-context.ghost.io)with homeless people suggesting she funds programs where people can use drugs. ![](https://storage.ghost.io/c/b3/f4/b3f40ad0-1485-4466-8b7d-5c45675dde69/content/images/2026/08/data-src-image-5f1f9f23-25f1-408a-88e2-a4fd72f0c5db.png) *Source:* [*Bruce Blakeman campaign*](https://x.com/NassauExec/status/2081733845812908447?ref=ai-in-context.ghost.io) That’s what I find most concerning – when AI is used to generate the likeness and voice of public figures without their permission and that has them saying and doing things they have not or would not, and portraying events that have never occurred. I’m not alone in this concern. A [Pew Research survey](https://www.pewresearch.org/short-reads/2024/09/19/concern-over-the-impact-of-ai-on-2024-presidential-campaign/?ref=ai-in-context.ghost.io) from 2024 finds that a majority of Americans, Republicans and Democrats alike, are concerned that AI will be used by campaigns to create and share false information that deceives the public. None of the Blakeman campaign ads sport the required disclosures that the New York state legislature passed. The [state requires ](https://www.nysenate.gov/legislation/laws/ELN/14-106?ref=ai-in-context.ghost.io)that political ads with “materially deceptive media,” including the use of AI, have clear disclosures that warn viewers that the content has been manipulated. The law extends to deceptive media that is used to recreate an actual person or to portray events that did not actually happen. Blakeman has defended his ads as [political satire](https://abc7ny.com/post/nassau-county-executive-bruce-blakeman-defends-use-ai-generated-campaign-ads-gov-hochul-calls-lying/19558586/?ref=ai-in-context.ghost.io), which is an exception carved out in the state’s disclosure laws. The New York Democratic Committee [submitted a complaint](https://www.thecentersquare.com/new%5Fyork/article%5F6a88065d-beaa-4298-a616-d791a045980e.html?ref=ai-in-context.ghost.io) to the New York Board of Elections challenging the ads as violating state campaign disclosure laws. Democratic Committee Chair Jay [Jacobs noted ](https://global.oup.com/academic/product/dirty-politics-9780195085532?cc=us&lang=en&&ref=ai-in-context.ghost.io)that Blakeman’s not in the entertainment business, so the political ads are fraudulent. Satire or not, when a political candidate’s campaign ads reproduce the likeness of an opponent and puppeteer them into saying and doing things they would not, the risk is that the public is misled. Research shows that the public [increasingly consumes their news from social media](https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026/dnr-executive-summary?ref=ai-in-context.ghost.io). The browsing habits on social media lead to accidental exposure to political content. A voter may only witness a few seconds of a video before moving to something else. Yet, what they are exposed to can be retained leading them to believe the false information and then even [sharing it online ](https://www.sciencedirect.com/org/science/article/abs/pii/S1066224322000557?ref=ai-in-context.ghost.io)to others. Even when people watch the full ad, they [likely won’t realize that it’s AI-generated](https://www.sciencedirect.com/science/article/pii/S2451958824001714?ref=ai-in-context.ghost.io). So, the consequence with ads like Blakeman’s is that people very likely may believe wrongly that Hochul or Mamdani made statements about their policies that they never said (or infer consequences of policies that are false).[\[2\]](#%5Fedn2) Our public sphere is already challenged by false and misleading claims about policies and politicians. AI use in politics just supercharges the problem even more. If the public can no longer trust what they see or hear, how are they to make informed decisions for themselves and for the common good? --- [\[1\]](#%5Fednref1) Kathleen Hall Jamieson wrote extensively about the misleading George H.W. Bush ads in her book [*Dirty Politics*](https://global.oup.com/academic/product/dirty-politics-9780195085532?cc=us&lang=en&&ref=ai-in-context.ghost.io). [\[2\]](#%5Fednref2) By the way, disclosures may not work either. Although the research is somewhat contradictory, one study finds that people may be [less trusting of content that is manipulated](https://pmc.ncbi.nlm.nih.gov/articles/PMC12166545/?ref=ai-in-context.ghost.io), but it doesn’t reduce sharing. Another study suggests that disclosures [don’t reduce people believing ](https://www.psypost.org/slapping-an-ai-generated-label-on-political-messages-doesnt-stop-people-from-believing-them/?ref=ai-in-context.ghost.io)the manipulated information. As is often the case: more research is needed here.