When I first started in law enforcement, many officers were still handwriting their police reports. Handwritten reports were often rushed, sloppy, and packed with misspellings and grammar mistakes. Around 2005, most United States’ law enforcement agencies started to use more advanced report management systems (RMS) to help write, edit, manage, and send police reports. These systems made report writing faster and a whole lot easier, if no less error-prone than handwritten reporting.
Many of today’s law enforcement agencies are transitioning to the third generation of RMS software. RMS 3.0 technology represents a significant improvement over earlier versions. This futuristic technology helps officers write and edit quicker by using pre-fill and auto-populate options that prevail among web and mobile applications today. More importantly, RMS also uses technology that is deeply integrated into other data mining and management software platforms, which allows for rapid data retrieval and real-time content verification.
The fascinating technology underlying RMS 3.0 uses deep learning algorithms, which are similar to Google’s general-purpose Artificial Intelligence (AI) and Machine Learning (ML) technology. AI and ML technology makes large scale data management possible. It also can be used to quickly and accurately find minute details hidden within an officer’s report or to flag potential high-liability issues, mistakes, or ambiguous statements.
Basic Format and Why It Matters
Understanding the raw format of a police report is vital to understanding how police agencies use AI and ML in police report writing. Because the early stages of ML can be temperamental, an officer must write a report in a specific way to allow the machine (program) to learn properly.
The basic structure of a police report includes:
Fact Sheet
This is the fill-in-the-blank type form used for generalized data: names, date of births, locations, type of offenses.
Narrative
The narrative is the most important section of any report. This is the section where the officer writes the when, where, what, why, and how of their investigation. This is also the section where AI and ML technology is used the most. Each narrative also includes subsections:
- Synopsis: A brief overview of what happened including dates, times, and location
- Pre-Arrival: The information of what the officer did before they arrived on scene
- Arrival: What the officer did when they arrived
- Investigation: What the officer did after they arrived
- Arrest and Charges: Specific details used to establish probable cause for charging
- Case disposition: Specific request for follow up, or request to close the case
- Attachments: Anything else relevant to the case (pictures, witness statements)
Because RMS 3.0 uses ML, an officer must use specific subheadings like Synopsis or Arrest and Charges or Interviews, to allow search crawlers to start linking data automatically. The more data (words, numbers ext.), the more information these crawlers can link. After the officer writes the narrative, they can use integrated AI technology to highlight inconsistencies, mark questionable information, and yes, flag poor grammar and spelling.
RMS 3.0 developers are currently working on a general-purpose AI that can also search an officer’s narrative for specific legal elements of a specific crime. These legal elements are used in the prosecution of a criminal case. If an officer misses an element, the prosecutor will dismiss the case, the bad guy goes free, and the law enforcement agency will likely be sued. AI technology quickly flags sentences or ask for clarification if something is missing so that all legal elements are met.
AI search crawlers use titles, phrases, headings, subheadings, and keywords to guide them. If this technology sounds familiar, it is because it is: private businesses have been using it for years.
How officers use SEO in police writing
Search engine optimization (SEO) was created to increase the visibility of websites, but this technology has inadvertently snuck its way into police RMS.
Using RMS 3.0 AI technology, an officer, supervisor, or analyst can use a Google-like search platform to search all reports for specific information. Data crawlers will then scan every document looking for key pieces of information. Before SEO integration, officers could only search fact sheets, not narratives or attachments.
Using this search platform, an officer can type: “white male, red shirt, tattoo on arm.” The system will then search all reports and documents for a white male wearing a red shirt with an arm tattoo.
How officer use attachments and scanned documents
The most important piece of AI and ML algorithm is its ability to search scanned documents and attachments. Witness statements, for example, are often briefly summarized in police narratives resulting in lost data—sometimes crucial data. The RMS uses optical character recognition technology that converts pictures or PDFs into words, allowing the search engine to scan all documents attached to the reports for specific information.
Why AI and ML matter
Officers hate to write reports. It is time-consuming, stress-inducing, and challenging. However, analyzing police reports to find usable, real-time data is far worse than actually writing a report—just ask your local police analyst.
RMS 3.0 technology directly integrates with up to a dozen internal police-related software programs and indirectly links to several hundred non-police related programs and data mining companies. Searching all these databases individually would be impossible, but using AI and ML is like using a strong magnet while looking for a needle in a haystack – it can quickly discard useless information to attract what’s pertinent.
Police technology, just like everything else, continues to evolve rapidly. Law enforcement agencies are finally using these AI and ML technology in police report writing. RMS ensure accuracy in police reporting and can improve officers efficiency, protect the innocent from being convicted, ensure rightful prosecution, accelerate data analysis and retrieval, help agencies flag issues, and potentially save lives.
