top of page

Using AI in Cell Phone Forensics

4 days ago
5 min read
Sample Executive Summary from AI Report
Sample Executive Summary from Cell Phone Forensic AI Report

AI is taking over our world in almost every aspect of our lives whether you know it or not. What about AI in the legal system? Does it help us or put us out of a job? Will there one day be an AI judicial system where we feed it all the case discovery and it spits out a verdict? No need for attorneys, judges, witnesses, experts?   Is it technically possible? Yes. Would it be more uniform across venues? Most likely. Would it be more correct than our current system? Maybe. Would it be acceptable by humanity? Highly doubtful but maybe some day in the future.


Let’s discuss what AI is and isn’t. I hear comments all the time like AI did this or AI did that. The thing that confuses people is what AI is and does compared to software code. Most processing and results are done by software with no need to access an AI model at all. For example, someone says: "Wow, my GPS used AI to find the fastest route." Usually, no AI was involved. Traditional software can read map data, traffic data and calculate the fastest route. Where AI could enhance that would be to add in your travel habits, the weather, where you usually stop for coffee and recommend a different route and departure time.


In cell phone forensics, software can be programmed to convert a raw AT&T CDR file into a Google Earth map. This does not require AI. However, if you want to add transcripts, interviews, cell phone data, case descriptions, etc. and tell it to compare claims and statements across all these data then that is using AI. AI adds reasoning and logic where software follows a specific set of instructions.


I have been using AI quite a bit lately, mostly doing software coding. I started with CoPilot and Gemini helping me create software utilities that converted files, streamlined processes, searched and sorted data. Using these utilities was quite helpful and saved many hours of my time as a cell phone data forensic expert.


I realized it could do more. I eventually purchased a Claude Pro subscription and started developing full-blown software programs that are easy to use and quickly convert raw data file into maps, reports, etc. This saved many hours of my time and my clients’ time. Multiply the savings I am getting by millions of people, and it is truly a generational jump in productivity gains. We are early in this endeavor. The more these AI models get used, the more people realize it can do. Simple things like re-organizing files and folders on my computer which might take an hour or two, AI can do in minutes, and do it better.


Of course, with any new technology there are bad actors and AI is no different. The bad actors won’t come from the technology developers, they will come from current bad actors that create malware, viruses, etc. AI needs to be monitored properly but should not be feared. It will take humanity to the next level. Countless new inventions since the caveman days have caused the same protests. Here is a bit of humor that addresses the issue. Cavemen Issue Dire Warning That The Wheel Is Going To Eliminate Countless Jobs


Let’s step back now to my field, cell phone forensics. How can AI assist the forensic examiner, attorney or investigator? To be useful AI must improve productivity by reducing time spent and focusing available time on high-priority tasks. AI chat bots are useful, can assist with research and provide knowledge quickly. However, if consistency and data quality is important then we must create a set of AI instructions that feed the model; a front-end into the model telling it exactly how to go about performing its reasoning specific to the task at hand.


In the area of cell phone forensics AI can be used to compare multiple datasets that include transcripts, interviews, police reports, witness statements, legal motions, presentations, photos, license cell phone network data, cell phone extraction data, cloud data (Facebook, iCloud, Backups, etc.), case descriptions, notes, case theories, etc.


In discovery the State will supply a mountain of digital files to the Defense and the attorneys are tasked to go through it all and make sense of it. AI can alleviate much of that burden, but you can’t just plug it all into a Chat bot. You need an organized software-based instruction set. The following shows the input screen of the AI Case Analyzer in WebMap.


WebMap AI Analyzer - Drag and drop files into AI model
WebMap AI Analyzer - Drag and drop files into AI model

The user drags in all the relevant case files plus the cell phone, Facebook, Cellebrite data, etc. The they click on “Find Claims” and the key parts of the data excluding any private information, is sent to the model. The model searches through the data and creates a list of “Claims” as shown in the following figure.


WebMap AI Analyzer - Identify and select Claims
WebMap AI Analyzer - Identify and select Claims

The attorney, investigator or forensic analyst can mark which claims it wants AI to research to generate its report. Simply click “Check the selected claims” and AI processes all the data.


The software contains a set of instructions that is sent to the model to provide a consistent and high-quality forensic output. You can’t just plug in information into ChatGPT and expect it to give you consistent answers from one case to another or even from one request to another. We have all seen how AI chat bots can make mistakes.


When the processing is complete AI generates a pdf and html version of the report. The report provides a detailed analysis of all the claims it found and identifies area of agreement, disagreement, consistent/inconsistent, items to explore further, other observations, etc. It compares phone data with claims from interviews and other documents and files. It can analyze if two phones most likely travelled together.  It cross compares claims from the inputted documents. It provides links to all the source records and document line numbers for backward source tracking.


I created a fictitious case for the purposes of demonstration. To read the report for this case click on this link.

.This report is based on a short and simple fictitious case. A typical case that includes inputted theories and notes and additional files will have many more claims.


While we are a long way from AI deciding our judicial fate, we can certainly use it to help add productivity improvements to the judicial system. With such a backlog of cases and long delays in trials this is a system that is prime for AI to help.

 
 
 

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating

© 2013 - 2026 by ZK Services, LLC. 

bottom of page