Incident Overview
In July 2026, three separate incidents in Florida and Washington demonstrated how AI-generated images and videos were used to fabricate crime reports and evidence, compromising the integrity of law enforcement processes. The cases involved a woman in St. Petersburg who submitted a ChatGPT-generated image of a nonexistent intruder to falsely report a burglary and sexual battery, a man in Lake Mary who used a deepfake video to claim a break-in of a deputy’s patrol vehicle, and Spokane police who circulated an AI-generated image of an injured dog as authentic before retracting it.
The incidents occurred between October 2025 and July 2026, with the St. Petersburg case involving Brooke Schinault, who pleaded no contest to misdemeanor false reporting in December 2025. Alexis Martínez-Arizala was arrested in Florida for the deepfake video incident, while Spokane police acknowledged the AI-generated dog image was fabricated after a supervisor mistakenly treated it as genuine.
These incidents highlight how AI-generated content can distort legal proceedings, waste emergency resources, and undermine public trust in law enforcement. The misuse of AI tools in these cases led to unnecessary investigations, misallocation of police and forensic resources, and reputational harm to affected agencies.
Petersburg case involved a ChatGPT-generated image resembling a viral AI ‘homeless man’ prank, while the Lake Mary incident used a three-second deepfake video shown to a deputy inside a store.
All three cases resulted in harm to epistemic integrity—the reliability of information—and, in the Florida incidents, caused wrongful strain on law enforcement and emergency responders.
Possible Root Cause
The misuse of AI-generated images and videos in these incidents stemmed from the accessibility of synthetic media tools and the lack of standardized verification protocols in law enforcement. ChatGPT and deepfake technologies were deployed by individuals to fabricate evidence, while Spokane police’s internal error revealed gaps in media authentication processes.
Petersburg case, Brooke Schinault used ChatGPT to create an image of a nonexistent assailant, which she submitted alongside false burglary and sexual battery reports. Police later discovered the image in a deleted folder dated before the alleged incident, linking it to a viral TikTok challenge. The fabricated report prompted officers, detectives, and forensic DNA collection efforts, diverting resources from genuine cases.
The Lake Mary incident involved Alexis Martínez-Arizala, who allegedly generated a deepfake video using AI tools to falsely claim a break-in of a deputy’s patrol vehicle. Surveillance footage later confirmed no such event occurred, yet the deputy initially treated the threat as real. This incident exposed vulnerabilities in how law enforcement assesses digital evidence, particularly when AI-generated content is indistinguishable from authentic footage.
Spokane police’s error occurred when an internally created AI-generated image of an undercover officer holding an injured dachshund was mistaken for an authentic photograph. A supervisor released the image to media after correcting an earlier claim that the dog had died.
The root causes include the ease of generating convincing synthetic media, the absence of AI literacy among some law enforcement personnel, and insufficient protocols for verifying digital evidence. Petersburg and Lake Mary cases involved malicious actors exploiting AI tools, while the Spokane incident highlighted systemic risks when agencies lack robust media authentication processes.
Takeaways
These incidents underscore the need for law enforcement agencies to implement rigorous verification procedures for digital evidence, particularly when AI-generated content is involved.
Key lessons include the importance of cross-referencing digital evidence with surveillance footage, forensic analysis, and witness statements before accepting claims as genuine. Agencies should also train personnel to recognize AI-generated content, as deepfakes and synthetic images increasingly resemble authentic media.
Petersburg and Lake Mary cases show how malicious actors can manipulate AI tools to fabricate crimes, leading to wrongful arrests, surveillance, and resource diversion. The Spokane incident reveals that even internal errors—such as jokes mistaken for real events—can result in public misinformation if not properly vetted.
To mitigate these risks, law enforcement should adopt AI detection tools, establish clear protocols for verifying digital evidence, and enhance training on synthetic media threats.
The harm caused by these incidents extends beyond individual cases. Misinformation from AI-generated evidence can lead to broader societal distrust in law enforcement, while wrongful investigations may result in wrongful arrests or surveillance of innocent individuals. Addressing these risks requires a combination of technical safeguards, procedural improvements, and public awareness campaigns.
The cases also reflect broader trends in AI misuse, where synthetic media tools are increasingly accessible to both malicious actors and well-intentioned but unprepared institutions. Without proactive measures, the integrity of legal and investigative processes remains at risk from fabricated digital evidence.
Sources
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1479/
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1450/
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1597/
Incident summary from the AI Incident Database, used under CC BY-SA 4.0.