This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
The Westbrook Police Department in Maine inadvertently shared an AI-altered photograph of drug evidence on Facebook after using an editing app to add their department patch. The AI tool significantly altered the packaging and attributes of the evidence, leading to public scrutiny and concerns about the reliability of police evidence. The department later confirmed the use of AI, apologized, and invited the media to verify the original evidence.
An evidence photo from a Westbrook, Maine, drug bust was reported to have been unintentionally altered by an AI‑powered editing tool before being posted to Facebook, changing packaging details and creating false visual elements. The altered image reportedly led to questions and concerns over the authenticity of police evidence. The Westbrook Police Department later confirmed the use of AI, explained the error occurred while adding their patch to the photograph, and issued a public apology.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
Also classified as
The police department relied on an unverified AI tool to edit official evidence and published the output without proper human verification.
The AI system failed to perform reliably by hallucinating and altering the background details of the evidence photo instead of only applying the requested patch.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by the AI system unexpectedly altering the evidence photo's attributes, which was an unintentional outcome of the user's goal to simply add a logo post-deployment.
A local police department in Maine inadvertently posted an AI-altered photo of drug evidence on social media after using an editing app. While the incident raised questions about the reliability of digital evidence in the legal system, it has negligible national security implications.
The AI system, ChatGPT, acted as an assistant within an editing app to help the user modify an image, requiring human oversight to verify the output before official use.
Risk Level 3: Limited Risk. The system involved is ChatGPT, a chatbot generating synthetic content, which falls under transparency obligations requiring disclosure of AI use.
The AI was used via an app to perform image retouching by adding a patch, though the underlying model is the ChatGPT chatbot.
The report clearly outlines the sequence of events: an officer used an AI-enabled app (ChatGPT) to add a patch to an evidence photo, which inadvertently altered other attributes of the image. The police department's public statements and subsequent apology provide direct, unambiguous evidence of the tool's role and the resulting public reaction.
None. The facts of the AI's unintended alteration and the police department's subsequent correction are clear and uncontested.
Technology
Unintended AI Image Alteration
The photo app's AI automatically modified evidence details without warning.
Generative AI Hallucinations
AI generated gibberish text and hazy edges instead of a clean edit.
Data Inputs
No Original Image Safeguards
The app processed the raw evidence photo directly without preserving original.
Human Factors
Lack of AI Tool Awareness
Officer did not realize the editing app used generative AI to alter images.
Confirmation Bias
Police assumed gibberish text was real packaging, delaying error detection.
Inadequate Visual Review
Staff failed to notice the obvious alterations in the image before posting.
Process and Methods
No Verification Protocol
No process existed to verify authenticity of modified images before release.
Deficient Social Media Review
The Facebook administrator posted the image without checking for edits.
Management
Lack of AI Guidelines
Management failed to provide policies on using AI tools for official media.
Poor Initial Incident Response
Management initially denied AI use before verifying the facts of the post.
Regulatory Environment
Lack of Evidence Integrity Rules
No clear rules govern the use of AI tools on digital forensic evidence.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
