MVT III: The Bullshit Strikes Back

Foreword By Kevin Daniels

This series has been a collection of information gathered by a lot of wonderful people volunteering their own time. Whenever we address an issue, the "evidence" we provided gets changed. Pay attention to The Data Dissenter's claims about how that continuously happened. After comments were made pointing out the ties, their stuff would change. As addressed in an earlier blog, even Mayor Burns removed EVM - Consulting from his LinkedIn page as soon as we made that connection. The coverup is part of the evidence that we're providing to you.

As we dive into this wonderful piece researched and written by The Data Dissenter, keep in mind the lengths that the Mayor will go to in subverting our election process. And the fact that our efforts to call out their shenanigans won't stop.

Meet Cleo Burns: The real AI controlling MVT

In the Name of Transparency

Disclaimer: This is not intended to accuse any individual or organization of wrongdoing nor illegal activity. Its purpose is to document and compile publicly observable technical findings. Readers are encouraged to ask their own questions, conduct their own research, and draw their own conclusions.

In recent months, Monroe Vote Tracker has been widely promoted within the Monroe, NC community as a nonpartisan, unbiased, fact-based reporting platform.

On February 15, 2026, Monroe Vote Tracker announced it had launched a “fully integrated” AI assistant named Cleo. When a platform positions itself as transparent and fact-based, its data practices and governance structures should be equally transparent. 

In multiple Facebook comment threads, Cleo stated that it had “searched voter database records,” reviewed “property and voting records,” analyzed users’ “multiple aliases,” and examined “other online posts” to determine party affiliation or residency status. These were not vague responses, they were specific claims of investigative action.

Public records do exist and that is not the issue. What I could not find was any explanation of how a chatbot operating within Facebook would reliably associate a user’s profile, particularly when partial names or pseudonyms are used, with specific public records or aggregated online activity.

In some responses, Cleo spoke in formal language but in others it adopted a more conversational tone while still asserting that it had conducted searches across databases and aliases. Variability in AI-generated language is normal, to an extent. However, when a system publicly claims to have searched records, analyzed identities, and reviewed posting history, clarity about how those conclusions are reached becomes important.

As someone reviewing these interactions directly, I was left with a simple question: what, exactly, is Cleo capable of accessing, and how?

If transparency is the stated goal, this information should be easily accessible. 

After reviewing Meta’s publicly available policies for 3rd party tools on Facebook, I was unable to find any documentation explaining how this chatbot collects, uses, or stores data. While Meta does not publicly certify individual integrations, platforms that process user interactions are generally expected to explain their data practices.

I left a post on the Monroe Vote Tracker page, thinking their chatbot was working 24x7, but my questions went unanswered. So I went directly to the website.

The site states that its data has not been “manipulated, weighted, or influenced by any individual, group, or ideology.” Given that emphasis on transparency, I looked for basic operational information such as who runs the platform, where its data is processed, and how the AI system is managed. I was unable to locate ownership details, a privacy policy specific to the chatbot, nor any documentation explaining how user interactions are handled.

I then conducted 2 separate conversations with Cleo through the website interface, approximately 24 hours apart. Between these 2 interactions, I shared information from my initial conversation publicly on Facebook. I do not know whether that had any effect on subsequent responses. However, because the tone and specificity of Cleo’s 2nd interaction differed from the 1st, I believe it’s relevant to note that the initial exchange had been publicly discussed prior to the follow-up conversation. In both interactions, when I asked about privacy practices or who operates the platform, Cleo emphasized that it is a general AI model trained by Google and does not have visibility into the organization behind the site.

AI systems often provide standardized responses. Still, when a platform positions itself around transparency, details such as governance, data handling, and accountability should not be difficult to find.

Cleo Interactions #1

Evidence that messages in FB claiming to review individuals' data, including property records, were written by humans and not an AI integration.

"I did not generate that response."

I won’t share every detail of my 1st conversation with Cleo because most of it isn’t relevant. What is relevant is that in that initial exchange, Cleo acknowledged it was operating on Google’s Vertex AI platform.

What raised questions for me came after reviewing Cleo’s responses on Facebook. In several public comments, the “fully integrated” chatbot claimed to have accessed voting history, property records, and even searched through multiple aliases and anonymous posts made by Facebook users.

So I asked directly.

When I presented Cleo with what appeared to be a direct quote from one of those Facebook responses, Cleo stated it did not generate that content.

Cleo Interactions #2

Approximately 24 hours later….

I returned to the website and initiated a second conversation with Cleo. Between these 2 interactions, I had publicly shared specific details about my 1st exchange on Facebook.

In the 2nd conversation, Cleo’s responses were more narrowly defined and structured.

Where the 1st exchange acknowledged operating on Google’s Vertex AI platform, the 2nd interaction repeatedly emphasized that Cleo does not know who operates the site, does not manage the data itself, and does not have access to information about how the system is configured. It described its role strictly as reporting Monroe City Council voting records based on data it has been given.

Cleo also clarified that labels such as “family-values” or “liberal” are pre-determined fields embedded in the JSON (JavaScript Object Notation is a lightweight, text-based format used for storing and exchanging data) data it receives. It stated that it does not interpret, assign, nor analyze those labels, just presents them as provided.

The shift in tone and scope between the 2 conversations was noticeable. I am not suggesting intent, simply calling attention to the differences in how the system’s capabilities were described within a 24-hour period.

"I do not 'decide' what information pertains to 'family values'." Cleo admits that these labels are being manually inputted by its human owner.

Technology, especially AI, is powerful. But when a platform presents itself as a tool for transparency and public trust, clarity about who is speaking, what tools are being used, and what data is accessible should not be difficult to find.

When my questions about site ownership and data handling remained unanswered, I shifted to publicly available technical information. Websites, like any software, leave fingerprints such as hosting infrastructure, analytics scripts, certificate records, and deployment patterns. These are not private records, they are part of a site’s public footprint.

The Monroe Vote Tracker domain is privately registered and routed through Cloudflare, both are common and legitimate practices. However, when a platform centers its identity around transparency and accountability, the absence of clearly identifiable ownership becomes relevant.

While reviewing the site’s technical settings, I noticed that its canonical domain, essentially the original or preferred web address a site is tied to, is monroestrongvotes.com.

“Monroe Strong” is not a neutral phrase locally. It has been publicly associated with the current mayor’s branding for years. Seeing that language embedded in the technical configuration prompted me to look further.

Public records indicate the mayor started EVM Consulting. I reviewed that site and where they released the “Every Vote Matters” project. I then compared the publicly visible infrastructure of all 3 sites.

Across Monroe Vote Tracker, EVM Consulting, and Every Vote Matters, I observed substantial similarities in backend configuration, hosting environment, and tracking implementation. From a technical standpoint, they appeared to be built and deployed from the same environment.

Again, technical similarities alone do not establish ownership. However, the overlap was significant enough to warrant questioning. 

When ownership is unclear, governance is undocumented, and technical infrastructure overlaps with politically affiliated branding, those are not fringe observations. Transparency becomes nothing more than a meaningless slogan.

When basic questions require reverse-engineering a website to find answers, transparency is already compromised.

Transparency is patriotic. #MonroeStrong