[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"cms-topics":3,"$f2jr6v5hyvnjva":34},[4,13,20,27],{"slug":5,"name":6,"intro":7,"introText":8,"seoTitle":6,"seoDescription":9,"essayCount":10,"indexable":11,"path":12},"trust-as-infrastructure","Trust as Infrastructure",null,"Signals, systems, networks, time, and resilience - and how they build on each other. The conventional wisdom on trust is wrong. This is the correction. Each article breaks down one assumption about authenticity, credibility, or signals... and rebuilds it from the ground up. Strategy for leaders who need credibility to actually hold.","The conventional wisdom on trust is wrong. This is the correction. Strategy for leaders who need credibility to actually hold.",18,true,"\u002Fwriting\u002Ftrust-as-infrastructure",{"slug":14,"name":15,"intro":7,"introText":16,"seoTitle":15,"seoDescription":17,"essayCount":18,"indexable":11,"path":19},"human-in-the-loop","Human in the Loop","AI won't replace your expertise. It will expose whether you had any. Personal essays from the professional middle: the judgment calls, identity questions, and uncomfortable moments that don't show up in the frameworks. For practitioners navigating the AI transition without a script.","AI won't replace your expertise. It will expose whether you had any. For practitioners navigating the AI transition without a script.",8,"\u002Fwriting\u002Fhuman-in-the-loop",{"slug":21,"name":22,"intro":7,"introText":23,"seoTitle":22,"seoDescription":24,"essayCount":25,"indexable":11,"path":26},"craft","Craft","The gap between what you intend to signal and what your audience actually receives is where trust dies. Practical reads on the visible layer: what your clients, readers, and critics see when they look, and how the thing got made that way. Because the strategy is only as good as what shows up on the other side.","The gap between what you intend to signal and what your audience actually receives is where trust dies. Practical reads on the visible layer, and how the work gets made.",12,"\u002Fwriting\u002Fcraft",{"slug":28,"name":29,"intro":7,"introText":30,"seoTitle":29,"seoDescription":31,"essayCount":32,"indexable":11,"path":33},"creative-technology","Creative Technology","Most of the interesting work happens at the seams between industries. The real-time engines built for games turn out to be what the water sector needed. Generative AI makes more sense as an org chart than as math. Notes from thirty years of pointing new tools at old problems, and watching what happens when someone aims them wrong.","Where new tools meet old problems. Real-time engines for the water sector, generative AI as an org chart, and how to aim technology at the right target.",5,"\u002Fwriting\u002Fcreative-technology",{"essay":35,"related":572},{"slug":36,"title":37,"subtitle":38,"excerpt":39,"publishAt":40,"topic":41,"heroId":42,"path":43,"body":44,"bodyText":267,"seoTitle":268,"seoDescription":269,"canonicalUrl":7,"dateUpdated":270,"citations":271},"deepfake-labels-wont-fix-election-ads","Deepfake labels won't fix election ads. Accountable sponsors might.","Every AI label answers how. Voters ask who sent it, and the law can only demand a name.","A label says how an ad was made. People decide what to believe by who sent it, which is why the sponsor's name has to cost something.","2026-11-05T13:01:00.000Z",{"slug":5,"name":6},"fedeec3b-883b-4eb4-97ae-ca320afdbf7c","\u002Fwriting\u002Ftrust-as-infrastructure\u002Fdeepfake-labels-wont-fix-election-ads",{"type":45,"children":46},"root",[47,55,60,65,70,77,82,87,92,97,102,108,113,118,134,140,145,151,156,161,175,198,203,209,214,219,224,233,238,243,257,262],{"type":48,"tag":49,"props":50,"children":51},"element","p",{},[52],{"type":53,"value":54},"text","A giant generative AI version of Jair Bolsonaro, Brazil's former president, told a São Paulo rally the truth about itself. \"What you see here is not me,\" the avatar said at the launch of his son Flávio's presidential bid, The Next Web reported. \"This is a simulation of my image and my voice using artificial intelligence.\" Then it campaigned for Flávio, in the name of a man serving twenty-seven years for plotting a coup, and a perfect label did nothing to slow it down.",{"type":48,"tag":49,"props":56,"children":57},{},[58],{"type":53,"value":59},"What it did draw was the who-question. Supreme Court Justice Alexandre de Moraes gave Bolsonaro forty-eight hours to say whether he had authorized the video, and raised a possible return to prison. His lawyers said he hadn't, which makes who was speaking an even sharper question.",{"type":48,"tag":49,"props":61,"children":62},{},[63],{"type":53,"value":64},"Ireland experienced the reverse. Days before its 2025 presidential vote, AI-generated RTÉ News bulletins falsely said Catherine Connolly had withdrawn from the election. An analyst told RTÉ that similar videos carried a \"synthetic\" label on YouTube and none on Facebook. The questions Ireland's electoral commission gave voters included \"Who is telling me this? Why are they telling me this?\"",{"type":48,"tag":49,"props":66,"children":67},{},[68],{"type":53,"value":69},"Every AI label answers one question: \"how was this made?\" Voters decide what to believe with a different one: \"who sent it?\" As The Next Web put it: \"A label tells you a video is fake. It does not tell you whether a jailed politician should be allowed to speak through it.\"",{"type":48,"tag":71,"props":72,"children":74},"h2",{"id":73},"the-label-that-mostly-didnt-show-up",[75],{"type":53,"value":76},"The label that mostly didn't show up",{"type":48,"tag":49,"props":78,"children":79},{},[80],{"type":53,"value":81},"The midterms were on 3 November, and the \"how\" label mostly sat this cycle out. By mid-October the Wesleyan Media Project had found at least 220 ads made or enhanced with AI, across thirty-four states. Its report doesn't break them down by office. News reports describe AI ads in a Kentucky House primary, by the losing candidate's account, and in Tennessee's governor's race.",{"type":48,"tag":49,"props":83,"children":84},{},[85],{"type":53,"value":86},"Only 27% carried a disclaimer about AI, and no federal rule requires one. That count includes uses beyond deceptive impersonations, so it's not a measurement of attempted deception. Spending behind them topped $137 million by Wesleyan's count, up from less than $50,000 in 2024.",{"type":48,"tag":49,"props":88,"children":89},{},[90],{"type":53,"value":91},"Where states legislate, they mostly choose the label. The National Conference of State Legislatures counted thirty-one states with election deepfake laws in June, and twenty-eight rely on disclosure. \"The fabricated ad can remain in circulation with that warning,\" writes Stefani Langehennig, an AI policy scholar at the University of Denver.",{"type":48,"tag":49,"props":93,"children":94},{},[95],{"type":53,"value":96},"Labels do have some impact. An American Association of Political Consultants Foundation brief, quoted by the Plain Dealer, called the disclaimer \"a cognitive speed bump\" that made viewers more skeptical. In a study Langehennig cites, labels lowered belief, but warnings that simply identified content as AI-generated \"did little to change how willing people said they were to engage with it, including sharing it.\" She notes the research didn't test whether disclosure laws change votes.",{"type":48,"tag":49,"props":98,"children":99},{},[100],{"type":53,"value":101},"The foundation's poll of 1,568 likely voters, reported by the Washington Post, shows what people judge. Around nine in ten called an AI video of an opponent saying something they never said \"completely unacceptable,\" regardless of whether it was disclosed. They'd made up their minds about the act.",{"type":48,"tag":71,"props":103,"children":105},{"id":104},"openly-ai-and-it-still-works",[106],{"type":53,"value":107},"Openly AI, and it still works",{"type":48,"tag":49,"props":109,"children":110},{},[111],{"type":53,"value":112},"Openly AI-generated content gets its power from whoever sends it. In December 2023, the opposition Pakistan Tehreek-e-Insaf (PTI) closed a virtual rally with a four-minute AI voice clone of its jailed figurehead, Imran Khan, captioned as the \"AI voice of Imran Khan based on his notes.\" AFP and Reuters reported more than 4.5 million views for the rally across Facebook, X and YouTube.",{"type":48,"tag":49,"props":114,"children":115},{},[116],{"type":53,"value":117},"In Turkey, the opposition Cumhuriyet Halk Partisi (CHP) gave its jailed Istanbul mayor, Ekrem İmamoğlu, an AI voice at a March 2025 rally in Maltepe, BirGün reported, without saying whether the video carried a label. The message said he wasn't afraid and wouldn't back down. Like Bolsonaro's avatar, these drew their authority from the jailed figure they spoke for.",{"type":48,"tag":49,"props":119,"children":120},{},[121,123,132],{"type":53,"value":122},"Fakes aim at the same signal. In Bangladesh's 2026 election, Global Voices reported on a study of seventy-two fact-checked AI cases, and 28% put false quotes in politicians' mouths through photocards mimicking trusted outlets. Seeing the quotes under several trusted brands, the piece's author writes, \"made the fabricated quotes more credible than isolated false claims.\" The RTÉ bulletins worked the same way. What they forged was ",{"type":48,"tag":124,"props":125,"children":129},"a",{"href":126,"rel":127},"https:\u002F\u002Fandrewmarconi.com\u002Fwriting\u002Fhuman-in-the-loop\u002Fbyline-floating-free-of-a-history",[128],"nofollow",[130],{"type":53,"value":131},"a newsroom's face",{"type":53,"value":133},".",{"type":48,"tag":135,"props":136,"children":139},"essay-gallery",{":items":137,"title":138},"[{\"asset\":\"d449a237-aeab-4675-aa23-48beafdb46f1\",\"alt\":\"A giant screen shows an AI avatar of Jair Bolsonaro above a crowd in Brazil's yellow team shirts, some filming on their phones.\",\"caption\":\"An avatar created by artificial intelligence simulated Jair Bolsonaro during a Liberal Party (PL) event and delivered a message of support for his son, Flávio, in the presidential race. Credit: PL\u002FYouTube.\",\"flags\":[{\"label\":\"Shows AI Generated Imagery\",\"color\":\"error\"}]},{\"asset\":\"61eae159-ea74-4104-815d-b4b063463369\",\"alt\":\"A deepfake RTÉ news presenter at a presidential election desk, beside a cut-out of a deepfake Catherine Connolly.\",\"caption\":\"The AI-generated videos featured deepfake versions of Catherine Connolly and RTÉ News presenter Sharon Ní Bheoláin. Credit: Raidió Teilifís Éireann (RTÉ).\",\"flags\":[{\"label\":\"Shows AI Generated Imagery\",\"color\":\"error\"}]},{\"asset\":\"2473b6b8-9799-4cb1-baea-a7481e7fabf7\",\"alt\":\"An AI-generated Imran Khan sits in a dark suit in front of a Pakistani flag, facing the camera.\",\"caption\":\"AI-generated image of PTI chief Imran Khan from a virtual party rally in December 2023. Credit: PTI\u002FYouTube.\",\"flags\":[{\"label\":\"Shows AI Generated Imagery\",\"color\":\"error\"}]},{\"asset\":\"c16a10ca-2b9d-4004-93fa-9b61c51c9b73\",\"alt\":\"Ekrem İmamoğlu in a grey suit speaks to camera in an ornate room, a Turkish flag and a large painting behind him.\",\"caption\":\"IMM President and CHP presidential candidate Ekrem İmamoğlu with a video voiced by artificial intelligence. Credit: ANKA.\",\"flags\":[{\"label\":\"Shows AI Generated Imagery\",\"color\":\"error\"}]},{\"asset\":\"9b464f42-c86c-46e5-b0c6-4159f75c78b2\",\"alt\":\"Two fake Bangladeshi news photocards side by side, in red, with Bengali quotes attributed to Tarique Rahman and Shafiqur Rahman under news outlet logos.\",\"caption\":\"Bangladesh: Fake photocards spreading misinformation targeting Prime Minister Tarique Rahman (left); fake photocards spreading misinformation targeting opposition leader and Jamaat Ameer Dr. Shafiqur Rahman (right). Credit: Dismislab.\",\"flags\":[{\"label\":\"Shows AI Generated Imagery\",\"color\":\"error\"}]}]","AI-generated political imagery from five countries",[],{"type":48,"tag":49,"props":141,"children":142},{},[143],{"type":53,"value":144},"The fake Biden robocall that hit New Hampshire in January 2024 forged two senders at once. Steve Kramer, a Democratic operative, cloned Biden's voice and spoofed the phone number of Kathy Sullivan, a former state Democratic official, CyberScoop reported. The FCC fined him $6 million, and a jury acquitted him of all the criminal charges in June 2025, the Associated Press reported.",{"type":48,"tag":71,"props":146,"children":148},{"id":147},"a-name-is-all-the-law-can-ask-for",[149],{"type":53,"value":150},"A name is all the law can ask for",{"type":48,"tag":49,"props":152,"children":153},{},[154],{"type":53,"value":155},"The law that asks the right question already exists. Ohio requires candidates and campaign committees to use the words \"paid for by\" followed by the name of the entity, and it bars a \"false, fictitious, or fraudulent\" name. Little in it makes the name tell you anything.",{"type":48,"tag":49,"props":157,"children":158},{},[159],{"type":53,"value":160},"Sometimes the line never arrives. The Ramaswamy campaign's ad featuring Amy Acton's estranged mother leaked without one, the Statehouse News Bureau reported, and Acton opponents and Ramaswamy supporters circulated it within hours, before the campaign confirmed it was theirs.",{"type":48,"tag":49,"props":162,"children":163},{},[164,166,173],{"type":53,"value":165},"Whether a leaked copy breaks the rule is ",{"type":48,"tag":124,"props":167,"children":170},{"href":168,"rel":169},"https:\u002F\u002Fpinkrosesociety.org\u002Fissues\u002Fdemocracy-voting-rights\u002Fmother-ad-paid-for-by-larose-complaints",[128],[171],{"type":53,"value":172},"an open question",{"type":53,"value":174},". Ohio Secretary of State Frank LaRose says his office is reviewing complaints about the ad, WSYX reported, and he wouldn't comment on the specifics. Straight Arrow News put it plainly while the sponsor was still unclaimed: \"Who is behind the video matters, both ethically and legally.\" The who-question applies however an ad was made.",{"type":48,"tag":49,"props":176,"children":177},{},[178,180,187,189,196],{"type":53,"value":179},"Sometimes the name arrives and says very little. In Ohio's governor's race, ",{"type":48,"tag":124,"props":181,"children":184},{"href":182,"rel":183},"https:\u002F\u002Fsignalohio.org\u002Fohio-insurers-utilities-drop-millions-on-governors-races\u002F",[128],[185],{"type":53,"value":186},"Signal Ohio reported",{"type":53,"value":188},", the Democratic Governors Association backs a group it calls Ohioans for Lower Costs, and the Republicans' offshoot is RGA Right Direction PAC, three letters and a slogan. ",{"type":48,"tag":124,"props":190,"children":193},{"href":191,"rel":192},"https:\u002F\u002Fandrewmarconi.com\u002Fwriting\u002Fhuman-in-the-loop\u002Fwas-it-generative-ai-ohio-governor-spin",[128],[194],{"type":53,"value":195},"One of Ohio's first AI ads",{"type":53,"value":197},", which put Sherrod Brown's face on an actor, came from the Ohio Flyer PAC, and Federal Election Commission records tie that PAC to Sen. Jon Husted's joint fundraising committee, the Plain Dealer reported.",{"type":48,"tag":49,"props":199,"children":200},{},[201],{"type":53,"value":202},"Europe shows where the line stops. The EU requires political ads to say who paid for them, among other rules, and Meta's answer was to stop running them there, citing limits on targeting and legal uncertainty. Politicians can still post organically: \"They just won't be able to amplify this through paid advertising.\" The sponsor line covers the paid slot, and sharing goes on without it.",{"type":48,"tag":71,"props":204,"children":206},{"id":205},"the-vouch-does-the-convincing",[207],{"type":53,"value":208},"The vouch does the convincing",{"type":48,"tag":49,"props":210,"children":211},{},[212],{"type":53,"value":213},"In a social feed, belief follows the person who shared it. In a Media Insight Project experiment with 1,489 American adults in late 2016, people saw a Facebook-style post of a diabetes story. It was shared by a public figure who often talks about health, one they'd already rated as trusted or not. Trust in the post was \"determined less by who creates the news than by who shares it.\"",{"type":48,"tag":49,"props":215,"children":216},{},[217],{"type":53,"value":218},"A trusted sharer with an unknown outlet beat a reputable outlet shared by someone untrusted, 49% to 32% on whether it got the facts right, and readers were readier to pass it on (39% to 23%). About half could recall who shared the story, but only about two in ten could name where it came from. That's news rather than ads, and it's ten years old.",{"type":48,"tag":49,"props":220,"children":221},{},[222],{"type":53,"value":223},"When people stop to check, they say they go to outlets they already trust. In the Reuters Institute's 2025 Digital News Report, 38% said they'd look there first, ahead of official sources (35%), fact-checkers (25%) and social media (14%). In the US, that meant each side's own: CNN on the left, Fox News on the right.",{"type":48,"tag":49,"props":225,"children":226},{},[227],{"type":48,"tag":228,"props":229,"children":232},"img",{"alt":230,"src":231},"Bar chart of the share who'd use each to check whether something was false, misleading or fake: a news source I trust 38%, official source 35%, search engine 33%, a fact-checking website 25%, somebody I know and trust personally 19%, comments from other users 18%, Wikipedia 17%, social media or video network 14%, an AI chatbot 9% (Reuters Institute, Digital News Report 2025)","\u002Fcms-assets\u002Fdc4dbc9d-dae9-428e-8719-4f9d31d70229?preset=hero",[],{"type":48,"tag":49,"props":234,"children":235},{},[236],{"type":53,"value":237},"The same channel carries bad sources just as smoothly. In July 2024, a fake video using a clone of Kamala Harris's voice went up on X labeled as a parody. Elon Musk reposted it eight hours later with \"This is amazing,\" a laughing emoji and no such label, PetaPixel reported, and NBC News put his post at more than 150 million views. Reactions split over whether it was parody, and neither report measures what viewers believed, but the audience came through the sharer and the label stayed behind.",{"type":48,"tag":49,"props":239,"children":240},{},[241],{"type":53,"value":242},"Jimmy Narang, describing his WhatsApp-style experiment in a UC Berkeley interview, says receivers \"update as if stories shared by their friends are ~66% likely to be true,\" while people share true and false stories at about equal rates, and about a quarter of stories get shared even when the sharer gives them less than a 50% chance of being true.",{"type":48,"tag":49,"props":244,"children":245},{},[246,248,255],{"type":53,"value":247},"So forwarding something doesn't equal vouching for it, regardless of whether it's true or false. I previously made ",{"type":48,"tag":124,"props":249,"children":252},{"href":250,"rel":251},"https:\u002F\u002Fandrewmarconi.com\u002Fwriting\u002Ftrust-as-infrastructure\u002Ftestimonials-are-dead-you-just-havent",[128],[253],{"type":53,"value":254},"this argument about testimonials",{"type":53,"value":256},": proof got cheap, and what still counts comes from someone with credibility at risk.",{"type":48,"tag":49,"props":258,"children":259},{},[260],{"type":53,"value":261},"If labels are going to be required, I think the \"who\" label is the one worth strengthening. An Election Commission of India advisory for the Bihar vote already joins the AI label to the name of the entity responsible, and Full Fact proposes one UK marker naming the promoter, on whose behalf it's published, and that it was altered. None of the research here shows that changes a vote. The sharer research points at the forward more than the sponsor, and the sponsor's name is the part of that chain a law can reach.",{"type":48,"tag":49,"props":263,"children":264},{},[265],{"type":53,"value":266},"A label tells you how an ad got made. You'll decide whether to believe it by who sent it, so the sender's name is the one that has to cost something.","A giant generative AI version of Jair Bolsonaro, Brazil's former president, told a São Paulo rally the truth about itself. \"What you see here is not me,\" the avatar said at the launch of his son Flávio's presidential bid, The Next Web reported. \"This is a simulation of my image and my voice using artificial intelligence.\" Then it campaigned for Flávio, in the name of a man serving twenty-seven years for plotting a coup, and a perfect label did nothing to slow it down. What it did draw was the who-question. Supreme Court Justice Alexandre de Moraes gave Bolsonaro forty-eight hours to say whether he had authorized the video, and raised a possible return to prison. His lawyers said he hadn't, which makes who was speaking an even sharper question. Ireland experienced the reverse. Days before its 2025 presidential vote, AI-generated RTÉ News bulletins falsely said Catherine Connolly had withdrawn from the election. An analyst told RTÉ that similar videos carried a \"synthetic\" label on YouTube and none on Facebook. The questions Ireland's electoral commission gave voters included \"Who is telling me this? Why are they telling me this?\" Every AI label answers one question: \"how was this made?\" Voters decide what to believe with a different one: \"who sent it?\" As The Next Web put it: \"A label tells you a video is fake. It does not tell you whether a jailed politician should be allowed to speak through it.\" The label that mostly didn't show up The midterms were on 3 November, and the \"how\" label mostly sat this cycle out. By mid-October the Wesleyan Media Project had found at least 220 ads made or enhanced with AI, across thirty-four states. Its report doesn't break them down by office. News reports describe AI ads in a Kentucky House primary, by the losing candidate's account, and in Tennessee's governor's race. Only 27% carried a disclaimer about AI, and no federal rule requires one. That count includes uses beyond deceptive impersonations, so it's not a measurement of attempted deception. Spending behind them topped $137 million by Wesleyan's count, up from less than $50,000 in 2024. Where states legislate, they mostly choose the label. The National Conference of State Legislatures counted thirty-one states with election deepfake laws in June, and twenty-eight rely on disclosure. \"The fabricated ad can remain in circulation with that warning,\" writes Stefani Langehennig, an AI policy scholar at the University of Denver. Labels do have some impact. An American Association of Political Consultants Foundation brief, quoted by the Plain Dealer, called the disclaimer \"a cognitive speed bump\" that made viewers more skeptical. In a study Langehennig cites, labels lowered belief, but warnings that simply identified content as AI-generated \"did little to change how willing people said they were to engage with it, including sharing it.\" She notes the research didn't test whether disclosure laws change votes. The foundation's poll of 1,568 likely voters, reported by the Washington Post, shows what people judge. Around nine in ten called an AI video of an opponent saying something they never said \"completely unacceptable,\" regardless of whether it was disclosed. They'd made up their minds about the act. Openly AI, and it still works Openly AI-generated content gets its power from whoever sends it. In December 2023, the opposition Pakistan Tehreek-e-Insaf (PTI) closed a virtual rally with a four-minute AI voice clone of its jailed figurehead, Imran Khan, captioned as the \"AI voice of Imran Khan based on his notes.\" AFP and Reuters reported more than 4.5 million views for the rally across Facebook, X and YouTube. In Turkey, the opposition Cumhuriyet Halk Partisi (CHP) gave its jailed Istanbul mayor, Ekrem İmamoğlu, an AI voice at a March 2025 rally in Maltepe, BirGün reported, without saying whether the video carried a label. The message said he wasn't afraid and wouldn't back down. Like Bolsonaro's avatar, these drew their authority from the jailed figure they spoke for. Fakes aim at the same signal. In Bangladesh's 2026 election, Global Voices reported on a study of seventy-two fact-checked AI cases, and 28% put false quotes in politicians' mouths through photocards mimicking trusted outlets. Seeing the quotes under several trusted brands, the piece's author writes, \"made the fabricated quotes more credible than isolated false claims.\" The RTÉ bulletins worked the same way. What they forged was a newsroom's face. The fake Biden robocall that hit New Hampshire in January 2024 forged two senders at once. Steve Kramer, a Democratic operative, cloned Biden's voice and spoofed the phone number of Kathy Sullivan, a former state Democratic official, CyberScoop reported. The FCC fined him $6 million, and a jury acquitted him of all the criminal charges in June 2025, the Associated Press reported. A name is all the law can ask for The law that asks the right question already exists. Ohio requires candidates and campaign committees to use the words \"paid for by\" followed by the name of the entity, and it bars a \"false, fictitious, or fraudulent\" name. Little in it makes the name tell you anything. Sometimes the line never arrives. The Ramaswamy campaign's ad featuring Amy Acton's estranged mother leaked without one, the Statehouse News Bureau reported, and Acton opponents and Ramaswamy supporters circulated it within hours, before the campaign confirmed it was theirs. Whether a leaked copy breaks the rule is an open question. Ohio Secretary of State Frank LaRose says his office is reviewing complaints about the ad, WSYX reported, and he wouldn't comment on the specifics. Straight Arrow News put it plainly while the sponsor was still unclaimed: \"Who is behind the video matters, both ethically and legally.\" The who-question applies however an ad was made. Sometimes the name arrives and says very little. In Ohio's governor's race, Signal Ohio reported , the Democratic Governors Association backs a group it calls Ohioans for Lower Costs, and the Republicans' offshoot is RGA Right Direction PAC, three letters and a slogan. One of Ohio's first AI ads, which put Sherrod Brown's face on an actor, came from the Ohio Flyer PAC, and Federal Election Commission records tie that PAC to Sen. Jon Husted's joint fundraising committee, the Plain Dealer reported. Europe shows where the line stops. The EU requires political ads to say who paid for them, among other rules, and Meta's answer was to stop running them there, citing limits on targeting and legal uncertainty. Politicians can still post organically: \"They just won't be able to amplify this through paid advertising.\" The sponsor line covers the paid slot, and sharing goes on without it. The vouch does the convincing In a social feed, belief follows the person who shared it. In a Media Insight Project experiment with 1,489 American adults in late 2016, people saw a Facebook-style post of a diabetes story. It was shared by a public figure who often talks about health, one they'd already rated as trusted or not. Trust in the post was \"determined less by who creates the news than by who shares it.\" A trusted sharer with an unknown outlet beat a reputable outlet shared by someone untrusted, 49% to 32% on whether it got the facts right, and readers were readier to pass it on (39% to 23%). About half could recall who shared the story, but only about two in ten could name where it came from. That's news rather than ads, and it's ten years old. When people stop to check, they say they go to outlets they already trust. In the Reuters Institute's 2025 Digital News Report, 38% said they'd look there first, ahead of official sources (35%), fact-checkers (25%) and social media (14%). In the US, that meant each side's own: CNN on the left, Fox News on the right. The same channel carries bad sources just as smoothly. In July 2024, a fake video using a clone of Kamala Harris's voice went up on X labeled as a parody. Elon Musk reposted it eight hours later with \"This is amazing,\" a laughing emoji and no such label, PetaPixel reported, and NBC News put his post at more than 150 million views. Reactions split over whether it was parody, and neither report measures what viewers believed, but the audience came through the sharer and the label stayed behind. Jimmy Narang, describing his WhatsApp-style experiment in a UC Berkeley interview, says receivers \"update as if stories shared by their friends are 66% likely to be true,\" while people share true and false stories at about equal rates, and about a quarter of stories get shared even when the sharer gives them less than a 50% chance of being true. So forwarding something doesn't equal vouching for it, regardless of whether it's true or false. I previously made this argument about testimonials: proof got cheap, and what still counts comes from someone with credibility at risk. If labels are going to be required, I think the \"who\" label is the one worth strengthening. An Election Commission of India advisory for the Bihar vote already joins the AI label to the name of the entity responsible, and Full Fact proposes one UK marker naming the promoter, on whose behalf it's published, and that it was altered. None of the research here shows that changes a vote. The sharer research points at the forward more than the sponsor, and the sponsor's name is the part of that chain a law can reach. A label tells you how an ad got made. You'll decide whether to believe it by who sent it, so the sender's name is the one that has to cost something.","Deepfake labels won't fix election ads. Sponsors might.","Most AI political ads carried no label, and a label only says how an ad was made. Voters decide what to believe by who sent it.","2026-11-03T23:04:32.180Z",[272,281,284,292,296,304,307,310,317,320,327,334,341,344,348,351,358,365,368,375,378,385,388,394,397,404,411,418,425,433,436,443,446,453,456,459,462,469,476,482,485,488,495,498,505,508,515,518,521,524,532,535,542,549,552,559,566,569],{"role":273,"relevance":274,"citation":275},"anchor","A labeled AI clip that campaigned in a jailed man's name.",{"rId":276,"title":277,"byline":278,"year":279,"url":280},"R5507","Bolsonaro is banned from speaking. His AI avatar campaigned anyway.","Alina Maria Stan",2026,"https:\u002F\u002Fthenextweb.com\u002Fnews\u002Fbolsonaro-ai-avatar-brazil-election-tse-deepfake-rules",{"role":273,"relevance":282,"citation":283},"States the how-versus-who gap in one line, from the Brazilian case.",{"rId":276,"title":277,"byline":278,"year":279,"url":280},{"role":273,"relevance":285,"citation":286},"Openly labeled AI that drew its pull from a jailed opposition leader.",{"rId":287,"title":288,"byline":289,"year":290,"url":291},"R5515","Ex-PM Khan uses AI-crafted speech to lure votes","AFP\u002FReuters",2023,"https:\u002F\u002Fwww.gulf-times.com\u002Farticle\u002F673834\u002Finternational\u002Fex-pm-khan-uses-ai-crafted-speech-to-lure-votes",{"role":293,"relevance":294,"citation":295},"context","Sizes the reach of the rally that closed with the labeled clip.",{"rId":287,"title":288,"byline":289,"year":290,"url":291},{"role":273,"relevance":297,"citation":298},"A fake that borrowed a trusted newsroom's face, faking who was speaking.",{"rId":299,"title":300,"byline":301,"year":302,"url":303},"R5503","AI-generated deepfake videos target presidential election","Kate McDonald",2025,"https:\u002F\u002Fwww.rte.ie\u002Fnews\u002F2025\u002F1022\u002F1539967-ai-generated-presidential-election\u002F",{"role":273,"relevance":305,"citation":306},"The how-label varied by platform, on one expert's account.",{"rId":299,"title":300,"byline":301,"year":302,"url":303},{"role":273,"relevance":308,"citation":309},"An election regulator's own advice includes the who-question.",{"rId":299,"title":300,"byline":301,"year":302,"url":303},{"role":273,"relevance":311,"citation":312},"Fakes that counterfeit the sender, across several parties.",{"rId":313,"title":314,"byline":315,"year":279,"url":316},"R5516","How artificial intelligence and synthetic reality shaped Bangladesh’s 2026 election","Zulker Naeen","https:\u002F\u002Fglobalvoices.org\u002F2026\u002F04\u002F01\u002Fhow-artificial-intelligence-and-synthetic-reality-shaped-bangladeshs-2026-election\u002F",{"role":273,"relevance":318,"citation":319},"Borrowed outlet brands, on the author's assessment; no audience measure is reported.",{"rId":313,"title":314,"byline":315,"year":279,"url":316},{"role":273,"relevance":321,"citation":322},"No federal rule asked these ads to disclose AI use.",{"rId":323,"title":324,"byline":325,"year":279,"url":326},"R5054","At Least $80 Million Spent on AI Ads This Cycle","Wesleyan Media Project","https:\u002F\u002Fmediaproject.wesleyan.edu\u002Freleases-090426\u002F",{"role":273,"relevance":328,"citation":329},"Voters judged the act; disclosure didn't change the verdict.",{"rId":330,"title":331,"byline":332,"year":279,"url":333},"R5482","AI ad mocks Talarico as Bible-defacing radical. Democrats bet it will backfire.","Matt Shuham","https:\u002F\u002Fwww.inquirer.com\u002Fpolitics\u002Fai-ads-political-campaigns-democrats-republicans-talarico-paxton-20261024.html",{"role":273,"relevance":335,"citation":336},"Most US deepfake law is a how-label.",{"rId":337,"title":338,"byline":339,"year":7,"url":340},"R5008","As AI political deepfakes go viral, lawmakers struggle to keep up – 2 states’ laws show limits of what they can do","Stefani Langehennig","https:\u002F\u002Fcobbcountycourier.com\u002F?p=112468",{"role":273,"relevance":342,"citation":343},"Under disclosure laws a fake meets the rule by carrying a warning.",{"rId":337,"title":338,"byline":339,"year":7,"url":340},{"role":345,"relevance":346,"citation":347},"counterpoint","Labels cut belief but left sharing largely alone.",{"rId":337,"title":338,"byline":339,"year":7,"url":340},{"role":273,"relevance":349,"citation":350},"The research stops short of voting.",{"rId":337,"title":338,"byline":339,"year":7,"url":340},{"role":345,"relevance":352,"citation":353},"Disclaimers do make viewers more skeptical.",{"rId":354,"title":355,"byline":356,"year":279,"url":357},"R1571","AI Political Ads Pop Up in Ohio Races as State Lawmakers Stall on Regulating Them","Jeremy Pelzer","https:\u002F\u002Fwww.cleveland.com\u002Fnews\u002F2026\u002F04\u002Fai-political-ads-pop-up-in-ohio-races-as-state-lawmakers-stall-on-regulating-them.html",{"role":273,"relevance":359,"citation":360},"The who-label already exists in Ohio statute.",{"rId":361,"title":362,"byline":363,"year":302,"url":364},"R5492","Section 3517.20 | Political communications; identification of source.","Ohio Revised Code (Legislative Service Commission)","https:\u002F\u002Fcodes.ohio.gov\u002Fohio-revised-code\u002Fsection-3517.20",{"role":273,"relevance":366,"citation":367},"The law can demand a real name; it can't make the name meaningful.",{"rId":361,"title":362,"byline":363,"year":302,"url":364},{"role":273,"relevance":369,"citation":370},"The who-label skipped; supporters carried the ad before the campaign confirmed it.",{"rId":371,"title":372,"byline":373,"year":279,"url":374},"R5495","GOP candidate for Ohio governor confirms ad featuring Democratic opponent's mother is his","Jo Ingles","https:\u002F\u002Fwww.statenews.org\u002Fgovernment-politics\u002F2026-10-20\u002Fad-featuring-mother-of-ohio-democratic-governor-candidate-sparks-controversy-and-backlash",{"role":273,"relevance":376,"citation":377},"Sharers carried the ad before the campaign confirmed it.",{"rId":371,"title":372,"byline":373,"year":279,"url":374},{"role":293,"relevance":379,"citation":380},"A newsroom framing the Ohio dispute as the who-question.",{"rId":381,"title":382,"byline":383,"year":279,"url":384},"R5496","Who's behind the video of an Ohio candidate's mother calling her a liar?","Straight Arrow News","https:\u002F\u002Fsan.com\u002Fcc\u002Fwhos-behind-the-video-of-an-ohio-candidates-mother-calling-her-a-liar\u002F",{"role":293,"relevance":386,"citation":387},"The PAC's name doesn't say who's behind it; FEC records do.",{"rId":354,"title":355,"byline":356,"year":279,"url":357},{"role":273,"relevance":389,"citation":390},"Committee names on both sides that say little about who's behind them.",{"rId":391,"title":392,"byline":393,"year":279,"url":182},"R5498","Insurers, utilities fund groups spending in Ohio governor’s race","Jake Zuckerman",{"role":273,"relevance":395,"citation":396},"A Republican committee's name: initials and a slogan.",{"rId":391,"title":392,"byline":393,"year":279,"url":182},{"role":273,"relevance":398,"citation":399},"The EU's who-label, separate from the AI Act's how-label.",{"rId":400,"title":401,"byline":402,"year":7,"url":403},"R5526","Transparency and targeting of political advertising","European Commission","https:\u002F\u002Fcommission.europa.eu\u002Fstrategy-and-policy\u002Fpolicies\u002Fjustice-and-fundamental-rights\u002Fdemocracy-eu-citizenship-anti-corruption\u002Fdemocracy-and-electoral-rights\u002Ftransparency-and-targeting-political-advertising_en",{"role":273,"relevance":405,"citation":406},"Meta left citing the regulation's limits; messages moved to sharing, where no sponsor line reaches.",{"rId":407,"title":408,"byline":409,"year":302,"url":410},"R5514","Ending Political, Electoral and Social Issue Advertising in the EU in Response to Incoming European Regulation","Meta","https:\u002F\u002Fabout.fb.com\u002Fnews\u002F2025\u002F07\u002Fending-political-electoral-and-social-issue-advertising-in-the-eu\u002F",{"role":273,"relevance":412,"citation":413},"A rule that joins the how-label to a who-label.",{"rId":414,"title":415,"byline":416,"year":7,"url":417},"R5513","EC Issues Fresh Guidelines on AI and Synthetic Content Use in Bihar Poll Campaigns","News On AIR (All India Radio)","https:\u002F\u002Fwww.newsonair.gov.in\u002Fec-issues-fresh-guidelines-on-ai-and-synthetic-content-use-in-bihar-poll-campaigns",{"role":273,"relevance":419,"citation":420},"A worked proposal for a stronger who-label.",{"rId":421,"title":422,"byline":423,"year":7,"url":424},"R5523","Second reading briefing – Representation of the People Bill: Establish stronger rules to deal with political deepfakes","Full Fact","https:\u002F\u002Ffullfact.org\u002Fdocuments\u002F410\u002FFull_Fact_briefing_-_political_deepfakes.pdf",{"role":273,"relevance":426,"citation":427},"In a feed, the sharer decides belief more than the publisher does.",{"rId":428,"title":429,"byline":430,"year":431,"url":432},"R5033","‘Who Shared It?’: How Americans Decide What News to Trust on Social Media","AP-NORC Center for Public Affairs Research and American Press Institute (Media Insight Project)",2017,"https:\u002F\u002Fapnorc.org\u002Fprojects\u002Fwho-shared-it-how-americans-decide-what-news-to-trust-on-social-media\u002F",{"role":273,"relevance":434,"citation":435},"The experimental result behind the who-sent-it claim.",{"rId":428,"title":429,"byline":430,"year":431,"url":432},{"role":273,"relevance":437,"citation":438},"Checking runs first through outlets people already trust; self-reported.",{"rId":439,"title":440,"byline":441,"year":7,"url":442},"R5058","Overview and key findings of the 2025 Digital News Report","Reuters Institute for the Study of Journalism, University of Oxford","https:\u002F\u002Freutersinstitute.politics.ox.ac.uk\u002Fdigital-news-report\u002F2025\u002Fdnr-executive-summary",{"role":293,"relevance":444,"citation":445},"US checking runs through each side's own trusted outlet.",{"rId":439,"title":440,"byline":441,"year":7,"url":442},{"role":273,"relevance":447,"citation":448},"A friend's forward works as a truth signal, for messaging apps outside the US feed.",{"rId":449,"title":450,"byline":451,"year":7,"url":452},"R5529","When Sharing isn't believing","UC Berkeley Economics","https:\u002F\u002Fecon.berkeley.edu\u002Fstories-interviews\u002Fwhen-sharing-isnt-believing",{"role":345,"relevance":454,"citation":455},"The vouch is only as good as the voucher; a share isn't a check.",{"rId":449,"title":450,"byline":451,"year":7,"url":452},{"role":293,"relevance":457,"citation":458},"Why he's barred from speaking, and whom the avatar spoke for.",{"rId":276,"title":277,"byline":278,"year":279,"url":280},{"role":273,"relevance":460,"citation":461},"The sender of a labeled clip was itself in dispute.",{"rId":276,"title":277,"byline":278,"year":279,"url":280},{"role":273,"relevance":463,"citation":464},"The court asked who was speaking, whatever the label said.",{"rId":465,"title":466,"byline":467,"year":279,"url":468},"R5634","Moraes dá 48h para defesa explicar vídeo de IA com Bolsonaro","Band","https:\u002F\u002Fwww.band.com.br\u002Fpolitica\u002Fnoticias\u002Fmoraes-da-48h-para-defesa-explicar-video-de-ia-com-bolsonaro-202607282301",{"role":273,"relevance":470,"citation":471},"The decision's own words on a possible return to prison.",{"rId":472,"title":473,"byline":474,"year":279,"url":475},"R5635","Moraes dá 48h para Bolsonaro dizer se autorizou vídeo de IA e cita possível volta à prisão","Agência Estado","https:\u002F\u002Fwww.opovo.com.br\u002Fnoticias\u002Fpolitica\u002F2026\u002F07\u002F29\u002Famp\u002Fmoraes-da-48h-para-bolsonaro-dizer-se-autorizou-video-de-ia-e-cita-possivel-volta-a-prisao.html",{"role":273,"relevance":477,"citation":478},"The latest count available before the election.",{"rId":479,"title":480,"byline":325,"year":279,"url":481},"R5631","GOP Lead in Senate Spending Narrows; Dems Catch Up in Senate Broadcast Airings","https:\u002F\u002Fmediaproject.wesleyan.edu\u002Freleases-101526\u002F",{"role":273,"relevance":483,"citation":484},"How rarely the how-label appeared across the cycle.",{"rId":479,"title":480,"byline":325,"year":279,"url":481},{"role":293,"relevance":486,"citation":487},"How fast AI-ad spending grew in one cycle.",{"rId":479,"title":480,"byline":325,"year":279,"url":481},{"role":293,"relevance":489,"citation":490},"A statewide race among the reported examples.",{"rId":491,"title":492,"byline":493,"year":279,"url":494},"R5633","Millions of dollars spent on AI-generated political ads as midterms approach","Charlotte Hazard","https:\u002F\u002Fweartv.com\u002Fnews\u002Fnation-world\u002Fmillions-of-dollars-spent-on-ai-generated-political-ads-as-midterms-approach-thomas-massie-ilhan-omar-zohran-mamdani",{"role":293,"relevance":496,"citation":497},"A US House race, on the candidate's own account.",{"rId":491,"title":492,"byline":493,"year":279,"url":494},{"role":273,"relevance":499,"citation":500},"Reporting from the day on the CHP rally.",{"rId":501,"title":502,"byline":503,"year":302,"url":504},"R5636","İmamoğlu, yapay zekayla hazırlanan videoyla Maltepe'deki yüz binlere seslendi","BirGün","https:\u002F\u002Fwww.birgun.net\u002Fhaber\u002Fimamoglu-yapay-zekayla-hazirlanan-videoyla-maltepe-deki-yuz-binlere-seslendi-611201",{"role":293,"relevance":506,"citation":507},"The message's words in Turkish; the paraphrase is ours.",{"rId":501,"title":502,"byline":503,"year":302,"url":504},{"role":273,"relevance":509,"citation":510},"How the experiment was set up.",{"rId":511,"title":512,"byline":513,"year":431,"url":514},"R5637","‘Who shared it?’ How Americans decide what news to trust on social media","American Press Institute","https:\u002F\u002Famericanpressinstitute.org\u002Ftrust-social-media\u002F",{"role":273,"relevance":516,"citation":517},"The size of the who-shared-it effect.",{"rId":511,"title":512,"byline":513,"year":431,"url":514},{"role":273,"relevance":519,"citation":520},"Trust in the sharer also moves the next forward.",{"rId":511,"title":512,"byline":513,"year":431,"url":514},{"role":273,"relevance":522,"citation":523},"People remember the sender more often than the outlet.",{"rId":511,"title":512,"byline":513,"year":431,"url":514},{"role":273,"relevance":525,"citation":526},"A fake that forged the sender as well as the voice.",{"rId":527,"title":528,"byline":529,"year":530,"url":531},"R5641","FCC hits operative behind New Hampshire robocall with $6 million fine","CyberScoop",2024,"https:\u002F\u002Fcyberscoop.com\u002Ffcc-fine-joe-biden-deepfake-new-hampshire-robocall-steve-kramer\u002F",{"role":293,"relevance":533,"citation":534},"The regulator's penalty.",{"rId":527,"title":528,"byline":529,"year":530,"url":531},{"role":345,"relevance":536,"citation":537},"The criminal case ended in acquittal.",{"rId":538,"title":539,"byline":540,"year":302,"url":541},"R5642","New Hampshire jury acquits consultant behind AI robocalls mimicking Biden on all charges","Holly Ramer","https:\u002F\u002Fwww.ksat.com\u002Fnews\u002Fpolitics\u002F2025\u002F06\u002F13\u002Fnew-hampshire-jury-acquits-consultant-behind-ai-robocalls-mimicking-biden-on-all-charges\u002F",{"role":273,"relevance":543,"citation":544},"The label stayed with the creator when the sharer passed it on.",{"rId":545,"title":546,"byline":547,"year":530,"url":548},"R5639","Elon Musk Shared Misleading AI Video of Vice President Kamala Harris on X","Pesala Bandara","https:\u002F\u002Fpetapixel.com\u002F2024\u002F07\u002F29\u002Felon-musk-shared-misleading-ai-video-of-kamala-harris-on-x-twitter-deepfake-election-presidential-usa",{"role":293,"relevance":550,"citation":551},"What the repost said in place of the label.",{"rId":545,"title":546,"byline":547,"year":530,"url":548},{"role":273,"relevance":553,"citation":554},"Sizes the reach that came through the sharer.",{"rId":555,"title":556,"byline":557,"year":530,"url":558},"R5640","How Elon Musk and a Kamala Harris deepfake ad sparked a debate about free speech and parody","Kat Tenbarge","https:\u002F\u002Fwww.nbcnews.com\u002Ftech\u002Fmisinformation\u002Fkamala-harris-deepfake-shared-musk-sparks-free-speech-debate-rcna164119",{"role":273,"relevance":560,"citation":561},"Why the missing line is still undecided.",{"rId":562,"title":563,"byline":564,"year":279,"url":565},"R5643","Ohio Secretary of State reviewing complaints over ad featuring Dr. Amy Acton's mother","Jeremy Edwards","https:\u002F\u002Fabc6onyourside.com\u002Fnews\u002Flocal\u002Fohio-secretary-of-state-reviewing-complaints-over-ad-featuring-dr-amy-actons-mother-frank-larose-election-integrity-commission-test-fine-transparency-case-political",{"role":345,"relevance":567,"citation":568},"People dispute whether it was parody.",{"rId":555,"title":556,"byline":557,"year":530,"url":558},{"role":293,"relevance":570,"citation":571},"Keeps the count from reading as a tally of fakes.",{"rId":337,"title":338,"byline":339,"year":7,"url":340},[573,582,591],{"slug":574,"title":575,"subtitle":576,"excerpt":577,"publishAt":578,"topic":579,"heroId":580,"path":581},"lawsuit-licensing-same-negotiation","The lawsuit and the licensing deal are the same negotiation","Answer engines stopped paying for attribution with clicks. Now it gets priced in court and in contracts.","AI summaries cut the clicks that paid publishers. Lawsuits and licensing deals now price attribution, and what you can prove you made is what gets paid for.","2026-10-20T12:00:00.000Z",{"slug":5,"name":6},"68a28f4e-4ad5-4db9-ae0a-0b6d0f031e6f","\u002Fwriting\u002Ftrust-as-infrastructure\u002Flawsuit-licensing-same-negotiation",{"slug":583,"title":584,"subtitle":585,"excerpt":586,"publishAt":587,"topic":588,"heroId":589,"path":590},"bias-meter-defended-klan","The Bias Meter Defended the Klan.","A newspaper automated the one thing it had left to sell, then found out what the machine said afterward.","The LA Times shipped an AI writing counterpoints faster than its editors could read. Oversight that can't keep pace with output is a backlog with a nicer name.","2026-09-29T12:00:00.000Z",{"slug":5,"name":6},"bd2d5eab-ba13-4e43-9121-c021e8cc72ab","\u002Fwriting\u002Ftrust-as-infrastructure\u002Fbias-meter-defended-klan",{"slug":592,"title":593,"subtitle":594,"excerpt":595,"publishAt":596,"topic":597,"heroId":598,"path":599},"news-trust-record-low","Trust in the news hit a record low. The least-trusted source is the one growing.","The fastest-growing news channel is the one people trust least, and your credibility has to survive the trip.","Trust in news fell to 37% while the least-trusted channel grew fastest. What survives an AI summary is receipts a machine can point at and people who vouch.","2026-09-22T12:00:00.000Z",{"slug":5,"name":6},"2b4ce95f-b276-4408-8313-750f44e4539f","\u002Fwriting\u002Ftrust-as-infrastructure\u002Fnews-trust-record-low"]