As campaign season gears up, AI-generated ads are everywhere

There is something almost quintessentially American about watching a campaign season adopt a shiny new technology with equal measures of enthusiasm, opportunism, and sheer nerve. The NPR report on AI-generated ads in the 2026 midterms captures a moment where political strategy and machine learning have fully converged. Candidates are no longer merely dabbling with the technology; they are building entire media operations around it, using generative models to produce ads that can be customized, tested, and deployed with breathtaking speed-knowledge that would have taken a traditional ad agency weeks to produce. Perhaps the most striking element is the asymmetry of perception. When a campaign uses AI, the same technology is framed either as innovative efficiency or manipulative deception depending on which side of the aisle is deploying it. Supporters see a candidate who is cutting-edge, responsive, and in touch with modern tools. Opponents see a manufactured image, an inauthentic construct designed to hoodwink voters. The reality, as with most political discourse, is that both perspectives contain a measure of truth. AI does allow for unprecedented personalization of campaign messaging, but it also makes it easier than ever to produce content that borders on disinformation. What gives one pause, however, is the quiet erosion of a shared baseline of political reality. When AI can generate a convincing image of a candidate doing something they never did, or produce a voice clone saying words they never spoke, the burden shifts onto the voter to determine the authenticity of everything they see. The report notes that data suggests many candidates are embracing AI to both boost their image and portray opponents unfavorably—a dual purpose that is entirely predictable but no less troubling. Campaigns have always distorted, exaggerated, and spun. The difference is that AI does so at industrial scale, with production costs near zero, and with a level of sophistication that makes fact-checking a losing game. Will it turn out voters? That remains the operative uncertainty. If history is any guide, voters may be more resilient than pundits fear reasonably skeptical of content they know could be synthetic, and perhaps even desensitized to it. Yet there is a deeper risk that goes beyond specific races. Every cycle in which AI-generated ads become more pervasive edges us closer to a state where trust in all political communication is collapsed entirely. And once that trust is gone, it is not easily rebuilt. The article, available at https://www.npr.org/2026/09/16/nx-s1-5953990/ai-political-ads-2026-midterms, stops short of declaring a verdict, and wisely so. The technology is too new, the data too incomplete. But the trajectory it outlines is one that every voter, regardless of party, should watch closely.