Seven days. That is how long a deepfake video falsely accusing a Taipei mayoral candidate of taking commissions on U.S. arms sales stayed online before platforms removed it, and that single number tells you almost everything about the state of AI and elections in November 2026.
In the span of 48 hours, election officials on three continents issued warnings, opened regulatory consultations, and demanded new laws aimed at synthetic political content. For campaign professionals, the message is unmistakable: AI is no longer a novelty in politics. It is the battlefield.
Why Did Taiwan's Election Commission Sound the Alarm?
Taiwan's Central Election Commission released a public service ad warning voters about AI-generated content ahead of the November 28 local elections, after deepfake videos targeting mayoral candidate Puma Shen circulated for a full week before removal. The delay demonstrated how quickly false claims can spread before takedowns occur.
According to the Taipei Times, Digital Affairs Minister Lin Yi-ching confirmed that authorities have asked Meta, Google, and Line to scan for improperly labeled AI content. Users who fail to label AI-created material could face legal consequences.
The case matters far beyond Taipei. When a fabricated claim about a candidate lives online for seven days, it reaches voters, gets screenshotted, and takes on a second life that no takedown can fully reverse. American campaigns heading into the 2028 cycle should treat this as a preview, not an anomaly.
What Is the EU Proposing for AI Election Content?
The European Commission opened a consultation on November 10 proposing that deepfakes be clearly labeled, retain those labels when shared, and that illegal deepfakes of candidates be removed quickly. The consultation closes November 23 and could reshape how platforms handle political synthetic media globally.
The proposed revision to election guidelines under the Digital Services Act zeroes in on a technical detail with huge practical consequences: labels must survive reposting. A disclosure that vanishes the moment a clip is reshared is no disclosure at all.
For political advertisers operating internationally, the signal is clear. Compliance is moving from a one time labeling decision to an ongoing content integrity obligation that follows every asset wherever it travels.
How Much Are Campaigns Actually Spending on AI Ads?
The Wesleyan Media Project identified at least $137 million spent on 220 AI-generated political ads this cycle, with 81% sponsored by Republican candidates or groups, according to a Boston Globe opinion piece published November 10. A separate roundup estimated roughly 300 AI-made or AI-enhanced ads backed by about $187 million.
The gap between those figures likely reflects different definitions of what counts as AI-generated, and that discrepancy matters before anyone cites either number. But both estimates point the same direction: synthetic creative is now a nine figure line item in American politics.
The partisan skew is striking but probably temporary. As tools get cheaper and case studies pile up, expect Democrats and independent groups to close the adoption gap fast. Campaigns that want the persuasive upside of AI without the reputational downside are investing in compliant, human-guided outreach instead, like the live voter conversations powered by HyperPhonebank, where real connection does the persuading.
Can Regulation Keep Pace With Synthetic Attack Ads?
Australia offers the latest test case. The Victorian Greens called this week for a ban on deceptive AI-generated political deepfakes and for truth in political advertising laws, after an AI-generated attack video involving the state opposition leader and the One Nation leader.
The party described the videos as "really dangerous when it comes to our democracy," and pointed to existing restrictions in South Australia and New South Wales as models. The pattern is consistent across democracies: a shocking synthetic ad appears, public outrage follows, and lawmakers scramble to catch up.
Smart campaigns are not waiting for regulators. They are building verification habits now, using tools like HyperPulse for synthetic polling and AI prediction that keep strategy grounded in modeled data rather than manipulated media.
Is There a Constructive Side to Election AI?
Yes, and it arrived the same week. Researchers at the University of Maryland unveiled VIOLETS, short for Voter Information from Official Local Election Trusted Sources, an AI chatbot trained on vetted information from Maryland and Montgomery County election offices.
The project aims to answer voter questions quickly and accurately while reassuring the public about election administration. It is a reminder that the same technology behind deepfakes can also route voters to authoritative answers, provided the system stays anchored to official sources.
That dual reality defines this moment. AI can corrupt an information environment or repair it, and the difference comes down to design choices, data sources, and human accountability. Campaigns exploring responsible deployment can find frameworks and training through the TPG Institute.
What Should Campaigns Do Right Now?
The lessons from November 10 and 11, 2026 are practical. First, assume any candidate can be deepfaked and build a rapid response protocol before it happens, because Taiwan's seven day takedown window shows how costly slow reactions are. Second, track the EU consultation, since its labeling persistence rules often become de facto global standards.
Third, audit your own creative. If your ads use AI assistance, document it, label it, and be ready to defend it publicly. The Wesleyan spending figures guarantee that journalists and watchdogs will keep digging into who spent what on synthetic content.
Finally, remember that voters still punish deception when they detect it, and they reward campaigns that treat them like adults. The winners of the next cycle will be the operations that pair AI's efficiency with old fashioned trust, and you can explore that balance through our campaign services or by following our campaign strategy coverage. The technology is moving fast. The campaigns that move faster, and more honestly, will be the ones still standing on election night.