It’s Not A Police Report, It’s A Transcription: A Better Term for AI-Generated Police Documents
Why police departments should avoid labeling AI transcription of body cam footage as report writing
Introduction
I propose that police departments should adopt a new terminology for the AI-generated documents based on body cam footage: scene transcription. This term reflects the nature and purpose of the transcription process, which is to provide a faithful and verbatim record of the speech and sounds captured by the body cam, without any interpretation, analysis, or opinion. Using the term scene transcription can help avoid the confusion and misunderstanding that may arise from labeling the AI-generated documents as reports. Reports are written accounts that require the writer's observation, analysis, and judgment, which are not present in the transcription process. In this white paper, I explain how scene transcription differs from report writing, why scene transcription is beneficial for police departments and the public, and how police departments can implement this change effectively[1][2].
Report Writing vs. Transcription
Report writing and transcription are two different processes that produce different types of documents. Report writing is the act of creating a written account of an event, situation, or problem, based on the writer's observation, analysis, and interpretation. Report writing requires the writer to select, organize, and synthesize relevant information, and to present it in a clear, concise, and coherent manner. Report writing also involves the writer's judgment, opinion, and perspective, which can be influenced by various factors, such as memory, bias, emotion, and context. Report writing is a complex and subjective task that requires critical thinking, creativity, and communication skills[3].
Transcription, on the other hand, is the act of converting speech into written text, without adding, omitting, or changing anything. Transcription requires the transcriber to accurately capture the words, sounds, and pauses of the speaker, and to use punctuation, capitalization, and formatting to indicate the structure and tone of the speech. Transcription does not involve the transcriber's interpretation, analysis, or opinion, but rather, it aims to preserve the original meaning and intention of the speaker. Transcription is a simple and objective task that requires attention to detail, accuracy, and speed[4].
Benefits of Using Scene Transcription
Using the term scene transcription for the AI-generated documents based on body cam footage can have several benefits for the police departments and the stakeholders. Some of the benefits are[5][6]:
· It can avoid the misconception that the AI-generated documents are the official, authoritative perspective of the officer, and comprehensive reports of the events captured by the body cam. This can prevent unrealistic expectations, false claims, and legal disputes.
· It can emphasize the fact that the AI-generated documents are only transcriptions of the speech and sounds recorded by the body cam, and not interpretations, analyses, or opinions of the police officers or the AI system. This can enhance the transparency, accountability, and trustworthiness of the AI system and the police officers.
· It can encourage the police officers to review, edit, and supplement the AI-generated documents with their own observations, analyses, and interpretations, and to clearly indicate the source and the nature of the information when necessary. This can improve the quality, accuracy, and completeness of the police documents.
· It can acknowledge the limitations and challenges of the AI system, such as speech recognition errors, background noise, speaker identification, dialects, slang, and jargon. This can motivate the police departments to monitor, evaluate, and improve the AI system and its performance.
· Acknowledges that an AI system might introduce inaccuracies in a transcript, indicating a certain level of precision, and attributes this to the documentation process rather than to any intention of an officer to affect the result.
· Allows for officers to have natural human interactive conversation with people that appears impersonal or simply an interaction that behaves in such a way to obtain proper documentation only.
· It can align with the existing terminology and standards of the transcription industry, which can facilitate the communication, collaboration, and compliance with the transcription service providers, the courts, and the public.
· This approach may assist departments in prioritizing how they allocate staff time, such as dedicating more resources to addressing unauthorized housing encampment connections rather than completing reports for speeding violations which could easily be documented from a transcription.
Best Practices for Implementing Scene Transcription
To successfully implement the term scene transcription for the AI-generated documents based on body cam footage, police departments should follow some best practices, such as[7][8]:
· Establishing a clear and consistent policy and procedure for using the AI system and the term scene transcription, and communicating it to the police officers, the transcription service providers, the courts, and the public.
· Categorizing various report types such as CAD notes, stops, incident notes, and probable cause affidavits.
· Cooperating with prosecutors and the public to distinguish between a transcript and a report, and ensuring that, if a report is produced, it's not created using AI prompts until AI use becomes more transparent and is widely accepted by both the legal system and society.
· Providing adequate training and guidance for the police officers on how to use the AI system, how to review, edit, and supplement the AI-generated documents, and how to distinguish between report writing and transcription.
· Creating a quality assurance and feedback mechanism for the AI system and the AI-generated documents, and ensuring that they meet the legal, ethical, and professional standards.
· Common tasks like neighborhood canvassing, checking speed signs, and calibrating radars should be recorded as separate free-form entry check boxes in the RMS rather than in the report narrative to prevent perjury risks.
· Monitoring and evaluating the impact and outcomes of using the AI system and the term scene transcription, and making adjustments and improvements as needed.
Conclusion
AI transcription of body cam footage can be a valuable tool for police departments, but it should not be confused with report writing. Report writing and transcription are two different processes that produce different types of documents. Police departments should use the term scene transcription instead of report writing for the AI-generated documents, as it can avoid confusion, misunderstanding, and legal challenges, and enhance transparency, accountability, and trustworthiness. To implement this change, police departments should follow some best practices, such as establishing a clear policy and procedure, providing adequate training and guidance, creating a quality assurance and feedback mechanism, and monitoring and evaluating the impact and outcomes.
[1] Forbes (2024) AXON AI Police Reports. Retrieved from https://www.forbes.com/sites/thomasbrewster/2024/04/23/axon-ai-police-reports-/
[2] Axon (n.d.) Draft One. Retrieved from https://www.axon.com/products/axon-records
[3] Feller, J. (2023) Legal Transcription vs Court Reporting: What’s the Difference? Retrieved from https://www.uslegalsupport.com/blog/legal-transcriptionist-vs-court-reporter/
[4] Transkriptor (2024) Comparing Dictation and Transcription: Key Differences and Benefits. Retrieved from https://transkriptor.com/dictation-vs-transcription/
[5] Bhati, N. (2023) AI Transcription: The Benefits, Applications, and Use Cases. Retrieved from https://appquipo.com/blog/ai-transcription/
[6] Krisp (2024) How to Transform Legal Transcription with the Power of AI. Retrieved from https://krisp.ai/blog/ai-legal-transcription/
[7] Canary, A. (2019) How to Transcribe Police Body Camera Footage For Better Law Enforcement. Retrieved from https://www.rev.com/blog/transcription-blog/transcribe-police-body-camera-footage
[8] Ramgahan, K. (2020) Improvement in Law Enforcement – Transcription of Police Body Camera Footage! Retrieved from https://gglot.com/improvement-in-law-enforcement-transcription-of-police-body-camera-footage/