it sure looks like David Kloiber is creeping on University of Kentucky employees to send them personalized mailers for the KY-6 primary

Kentucky primaries for the 2026 elections take place a week from today, so it’s not surprising that we’ve been getting some political mail over the past couple of weeks. Today, though, something came in the mail that really took me aback. David Kloiber’s campaign sent us something that was clearly more than a regular mailer, since it came in a letter-style envelope and was addressed to both me and my spouse.

Both of us are employees at the University of Kentucky, a public institution, so there’s a fair amount of publicly available information about us floating in the ether. That’s to be expected, but it doesn’t mean that someone can’t be creepy about it. The first paragraph of the letter that we got addressed one of us by name, listed their exact employer within the University, and had some personalized text in there that was clearly a riff on a publicly listed position title before pivoting into a “you work for this kind of institution, so this is why you should vote for me” spiel.

In response to Kloiber’s speaking directly to us, I’d like to speak directly to him for a paragraph: This is weird, my dude. Super weird. I don’t know who on your campaign thought this would make your outreach feel more personalized and be more effective, but you should see the number of 🤢 emojis, “creeeeppyyy"s, and “maybe I would have expected this from a GOP mailer"s that we’ve texted each other in the 45 minutes since I checked the mail today. This doesn’t say “I know who you are and I care about you” so much as “I looked you up and want to use what I learned for my personal benefit.”
Of course, it arguably gets better from here. I do a bit of web scraping myself, so there’s a (small) part of me that’s technically impressed with how this might have all worked. Here’s my best guess: Scrape the hell out of the UK website to identify employees, compare that list of employees with public or purchased voter info to identify folks likely to participate in a Dem primary, and then develop personalized mailers for all of them. Good so far, but: Even assuming that you can script the first two steps of that (which doesn’t strike me as terribly hard, even if I don’t know if I could do it easily myself), the level of personalization that’s in here (again, specific riffs off of specific job descriptions) is not something that would be easy to automate.

At least, it wouldn’t have been before 2023 or so. I’m pretty confident that there’s some LLM work going on here, and the awkward phrasing of some of the personalized language only makes me more confident. So now I wonder how much of this is good old fashioned web scraping (which, to be clear, can be plenty creepy on its own) and how much of this is a carefully massaged LLM prompt. Because nothing makes the appropriation of publicly available information for private gain (this mailer) worse than using a tool that is fundamentally built on the appropriation of publicly available information for private gain. Bravo.

Anyway, I’m going to be honest: I didn’t vote for David Kloiber in 2022 because he struck me as someone with more money than ideas. That impression has stuck enough in my head that although I haven’t done all my primary homework yet, I don’t know that I was all that likely to vote for him in the first place. That said, I take pride in being an informed voter, so I would genuinely have tried to read up on his policies and give him a fair evaluation before next Tuesday. Not now, though—I am not very interested in reading up on a candidate who thinks he can win my vote by creeping on me online and feeding the results into an LLM.

• 2026 elections
• David Kloiber
• generative Ai
• University of Kentucky
• privacy
• data privacy
• web scraping
• digital labor