![[AI "Who should I vote for?" cover.png]] We've been on the [[On Political Botslop|botslop]] beat for a minute. As 'AI' discourse infects every aspect of professional knowledge work, we think a thought from last year holds up pretty well: ![[On Political Botslop#^5d1ae8]] We were talking about content creation - and we think its holding up. As ever, AI generated or not, audiences hate [[An Evolving Collection of Politicians Pointing at Jobs and Leadership|fake stuff]], machine generated or not. [(It's rapidly getting worse](https://www.nytimes.com/interactive/2026/02/13/style/slopulism-trump-conservative-social-media.html)) But what about when the machine as "answer bot," to which users begin outsourcing their decision making on all manner of topics? The value of para-social relationships with machines is another one for the philosophers. Count us, [with the Holy Father,](http://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html) deeply skeptical at best. But they're here, and they're going to be tasked with millions of thorny questions of personal political judgment. Why think like a citizen when Claude will do it for you? As ever, our friend [[Eric Wilson]]'s latest summary in his must-read Campaign Trend newsletter is not to be missed. [Credit to him](https://campaigntrend.com/chatbots-get-their-election-test/) on how AI companies for a fantastic overview on how AI companies are handling election related inquiries. So far: think Google Search on street-legal steroids. Google is quickly redesigning their search function to AI, (including their advertising packages). And consumers are adopting in real time. Search is no longer a listing of top hyperlinks to outside web addresses, usually reported news sources; but a summary of those links, arranged in narrative format, sometimes with less obvious links to "supporting data" embedded. It isn't exactly revolutionary, yet. And sometimes it's dead wrong. [The hype is hype-y.](https://anchorchange.substack.com/p/voters-are-using-ai-to-fact-check?ref=campaigntrend.com) But it's a sea change for those managing perception. Try it now: ask Google about the upcoming election in Congress in your district, and see what Gemini serves up. This digital real estate is going to matter _a lot_ in the next five months. Cue the dawn of political actors gaming the system, before it all changes. The pace on that is TBD: fast, or faster. But it will be a profitable dark art. ### What's a proper, if less likely, future? So far, most chatbots act like Citizens Voter Guides: fair enough (pardon the pun). As the tech evolves, maybe the AI companies could share, in equal measure, what the campaigns say about themselves. Instead of trying to render Google Search Plus 1 - summarizing a few top websites and squeezing them through an LLM output - present human/visual arguments. Let the campaigns portray themselves as their investments show how they want to be portrayed? Render the most played YouTube ads from each campaign for a viewer, and let them hash it out? That's how humans [[Humans Learn Visually|internalize information]]. ![[AI Questions.png]] AI bots shouldn’t emulate the information and economic architecture we're all familiar with from two decades of searching the internet. That is: the top bidder wins the top spot in the race for attention. As advertisers, we want access to eyeballs to make a better case than the other guy. We don’t want a market where it’s a race to simply box out the other guy and block his access to those eyeballs. The model of the :30 second broadcast ad is fundamentally built on equal access to the information relayed. Each side gets space at the lowest rate in the market. The most fervent supporters fund the ability to move that message to voters. If one side has more supporters (money), or more personal will (money), an advantage can be built. Yes, it was and is a competitive attention market where money can, to a point, be predictive of outcomes. But in that traditional model; no one is bidding up for a *monopoly* on the attention of *one voter in one place* at "point of sale": theoretically, ready to make one of their two essential political decisions: whether to vote, and whom to support. And that's to say nothing of monopolizing attention using tools controlled by a single (or a couple) huge, government-intertwined companies. The AI in the pocket problem gets directly to why every state bans advertising and electioneering inside polling places. Now, a tidily organized summary of all that electioneering activity goes with you, right behind the voting curtain. Talk about a point of sale advertising opportunity. This on top of how [[First - Get people to watch|unwatchable ads get skipped]]. Boring content gets ignored by algorithms, political or otherwise. [[Private vs public attention]] distinctions are melting away. ##### Voters by the thousands - millions? - are going to ask chatbots whom to vote for standing in the voting booth this fall. We know we must adapt political messaging to ad inventories designed for commercial use. That's how the 30 second spot became the _de facto_ repository of American political argumentation. Nestle and Kellog's and Buick had three of four spots in a 2 minute commercial break. Senator So and So gets the fourth at the cheapest available rate. He got to make his case. Senator Who and Who had the same chance, same rate. USPS delivered both parties' mailers right alongside the Sears catalogs. Again, the fairness doctrine, however dated now, came from a valid civic impulse. The rules fit the medium. Political campaigns adapted to fit the commercial space. Any technology with so much power to sway voters making a choice in the moment shouldn’t be strict, winner take-all-bidding at immediate point of sale. It shouldn’t be easy to game. How about a base fee for candidates, at the scale of their races, to be included in chat bot summaries? *And require end users to consume ads from both sides*. Otherwise it'll be another battle for top-click real estate, which amounts to a race to the bottom. Users are going to these technologies to be told what to think based on personalized inputs; not simply to have information presented to them when their attention is forcibly interrupted from the ball game or the game show ([[Everything Is Television]]) Call it the fairness doctrine adapted to the 21st century. Let the best creative and the best arguments win out, when voters ask for political information directly. Use technological controls to *force* them to consider both sides. Which, we hasten to mention, is not the same thing as running ads in the chatbot stream of a voter asking for **other information.** If you’re asking Chat GPT for golf or hunting tips, and we’re in a GOP primary in your district, you can bet we want access to your eyeballs and should be able to pay top dollar for it in a competitive environment. Business is still business ;) Tech giants might improve their standing with politicians and with voters by balancing competitive argumentation vs. commercial maximization. --- PS - if your clients want to [First - Get people to watch], then understand [[Every Image Matters (The Power of the Frame)|the power of frame]], and win in the [[Attention Economy]].... well, we know some guys.