08/28/2026
Ancapistan update for Morgan County. Police are using chats in their jobs:
Yes. Police are using chat-style AI in three different ways. They are not the same thing.
1. Official police chat tools
These are vendor products built as chat or prompt interfaces for cops.
• Axon Draft One turns body-cam audio into a draft police report. Axon says it is in about 600 departments and has helped write 600,000 reports. It is built on an OpenAI GPT-4 variant. The officer is supposed to edit and sign off.
• Flock OS Investigate is an investigative chat box. Officers pick from 69 prewritten prompts or write their own, then the system searches plates, movement patterns, case files, and commercial identity data. WIRED reconstructed the code. Flock says it is still in development with a small group of agencies.
• Mark43 ReportAI / BriefAI drafts reports and summarizes cases from dispatch records and body-cam footage. Mark43 says it serves 300+ agencies.
This is the “approved chat” path: a department pays a vendor and treats the chatbot as part of the official stack.
2. Informal / “shadow” chats
Officers also use public ChatGPT, Copilot, and similar tools on their own — sometimes against policy.
Documented examples:
• A Bluffdale, Utah officer consulted ChatGPT on scene to decide whether hiring a private investigator violated a protective order. Department policy allowed AI; critics said using it for legal analysis was inappropriate.
• A federal judge noted an ICE / Border Patrol agent using ChatGPT to write a use-of-force narrative from a short description plus photos. The judge said that may help explain inaccurate reports.
• Police1 now warns departments about “shadow AI”: officers pasting case facts into public ChatGPT to draft reports or summarize interviews, outside IT control.
• In the UK, officers have been investigated or fired for unauthorized ChatGPT use on police work. That is overseas, but it shows the same pattern.
3. Police using other people’s chats as evidence
This is the reverse: investigators pulling civilian ChatGPT / Gemini logs as evidence of intent. That is already happening in U.S. cases. It is not police using a chat to do the work. It is police treating a chat log like a diary.
Local relevance
I have no public confirmation that Hartselle PD is using ChatGPT or Draft One.
What is confirmed nearby:
• Decatur just approved a large Axon package (body cams, vehicle cams, mobile ALPRs, drones, AI software). Draft One is Axon’s flagship report chatbot, so it is the most likely official chat tool if Decatur turns that feature on. That is inference from the contract family, not a published “Decatur uses Draft One” statement.
• Hartselle’s new $300k upgrade is cameras and radios. It does not, by itself, prove a chatbot rollout.
The practical split
Type
What it is
Status
Official vendor chat
Draft One, Flock OS Investigate, Mark43
Growing fast, paid, department-controlled
Shadow ChatGPT
Officer uses public chatbot on a phone or computer
Happening, often poorly controlled
Civilian chat logs as evidence
Police seize or receive chatbot transcripts
Already in court files
So yes: police are using chats. The official version writes reports and runs investigative prompts. The unofficial version is officers asking ChatGPT legal or case questions. The third version is police reading your chats.
If you want the next cut, I can draft the exact public-records questions to ask Hartselle PD and Decatur PD: which AI chat tools they authorize, whether officers may use public ChatGPT, and whether any reports have already been AI-drafted.
🦅 ANCAPISTAN:�🟢 STANDING BY — Daily briefing available
