Networks of hyperlocal websites are publishing political content designed to shape what AI chatbots say about US candidates, according to CyberScoop. The reported operation spans at least 14 states and 460 pages, turning election influence towards the answers people get from AI, not just the posts they scroll past.
Watch Desk analysis
What happened
CyberScoop reports that the sites present themselves as local news or opinion outlets, cite one another, and repeat short endorsements of candidates on issues such as housing and taxes. VisionHeight, the threat-intelligence firm behind the research, says the content appears designed for AI crawlers rather than human readers, with major US AI companies allowed to scrape the sites.
The report says VisionHeight first identified 37 websites across six states, then found the operation reached at least 14 states and 460 pages. It does not attribute the activity to a particular actor. CyberScoop says researchers confirmed at least one major AI model, Perplexity, cited one of the pages as its first source in an answer. That is evidence of a citation, not proof that the operation changed an election result or broadly shifted chatbot responses.
Why it matters
The tactic targets a new information middleman. Rather than persuading voters to visit a propaganda site, an operator can try to get a chatbot to retrieve its candidate-friendly material when someone asks a local political question. A network of sites citing one another may look like a small local press corps to a crawler. It is rather less reassuring when the newsroom is a template.
The reported scale is also striking: CyberScoop says roughly 100,000 words of candidate content appeared on the network on 30 September, and 36 of the 37 initially identified sites were registered within an eight-minute period. That suggests the infrastructure can be assembled quickly, though the report does not identify who built it.
Our read
This is a substantive warning about how AI search and retrieval can be gamed, not proof that a chatbot has been captured or that voters were persuaded. The practical distinction matters: a cited page can surface in an answer without establishing that the answer, or anyone’s vote, changed. AI providers should explain how they assess clusters of sites that manufacture local credibility, and show whether those pages are influencing answers.
What to watch
- Whether VisionHeight or other researchers publish further technical details about the network and its reach.
- Whether AI providers change how they assess sources that cite one another in coordinated networks.
- Whether researchers find the same tactic affecting chatbot answers beyond the single citation described by CyberScoop.
Discussion spark: If a chatbot cites a coordinated network of fake local sites, should the provider be responsible for detecting and labelling it, or should that burden fall to users and independent researchers?
Sources and evidence
- Source update (15 October 2026, 19:29 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.