The Heber City Police Department in Utah got more than they bargained for when testing new AI report-writing software in December 2025.
Instead of generating a standard incident summary, the artificial intelligence decided to document something far more magical: an officer transforming into a frog.
The embarrassing glitch happened during a mock traffic stop designed to showcase Draft One, an AI tool from police tech company Axon that uses OpenAI’s GPT language models to automatically write reports from body camera footage.
The problem? The software picked up audio from The Princess and the Frog playing somewhere in the background and incorporated Disney dialogue into an official police document.
“The AI report writing software picked up on the movie that was playing, which happened to be The Princess and the Frog“, Sgt. Keel explained to FOX 13 News. “That’s when we learned the importance of correcting these AI-generated reports“.
When tech meets… Disney
Heber City PD was testing two different platforms that month: Draft One and Code Four, the latter created by MIT dropouts George Cheng and Dylan Nguyen, both just 19 years old.
Both systems promise to save officers hours of paperwork by analyzing body cam audio and generating full reports automatically.
According to Sgt. Keel, despite the amphibian incident, the technology actually works, saving him six to eight hours weekly once you account for the human oversight required.
The frog fiasco might sound like a one-off comedy sketch, but it highlights real growing pains as police departments rush to adopt AI.
In 2025 alone, AI security software flagged a student’s bag of Doritos as a firearm, and another system got a 99% match that led to an innocent man being arrested after being misidentified as a banned casino patron.
The human element still matters
Sgt. Keel admits he’s “not the most tech-savvy person”, but finds the software user-friendly and genuinely helpful for cutting down administrative work.
The catch? Every AI-generated report still needs human review before it can be considered remotely usable, something the department learned the hard way when their test document came back filled with fairy tale nonsense instead of traffic stop details.
As law enforcement agencies lean harder into automation for everything from report drafting to facial recognition, incidents like this serve as a reminder that the technology isn’t foolproof.
Human oversight remains essential, especially when Disney movies are involved.
Want more wild tech stories and geek news? Follow Geek Realm Hub on Facebook and join our community—we promise we won’t turn you into a frog!