This article presents the development of an Android application prototype designed to assist Military Police officers with writing incident reports. The application leverages Google's Gemini AI API to provide real-time grammar correction and intelligent suggestions for writing improvement, acting as an experienced officer reviewing and enhancing the text. The app's architecture, based on the MVVM (Model-View-ViewModel) pattern, and the user interface design, focused on simplicity and usability, are detailed. The implementation of the Gemini AI API, including the system prompt that defines the AI's role and behavior, is explored, along with cost analysis and comparison with other popular natural language processing APIs like OpenAI and Anthropic. Test results using real incident reports demonstrate the app's ability to improve text clarity, conciseness, and professionalism, as well as provide relevant suggestions for police practice, such as missing information, legal procedures, and evidence collection. The article discusses the impacts and benefits of the application for the PMPR, the limitations of the technology, and suggestions for future work, including integration with other platforms and expansion of AI functionalities.
Read full abstract