Crime 2.0: AI for Law Enforcement
I led the research and design behind an AI tool that puts pedophiles behind bars
As a member of a three-person team, I took on the responsibility of understanding the email analysis process for child pornography investigations and used that information to design and prototype the future of digital data investigation.
A tool that documents a narrative of criminal lifestyle
To aid prosecutors in putting criminals behind bars, agents need to compile digital evidence of illegal activity over extended periods of time. The original request to build a tool to expedite email analysis evolved into something more collaborative that solved the ultimate underlying issue of holding criminals accountable.
Inspired by Outlook, this tool supercharged investigative analysis with AI NLP & OCR features.
Background
What if we streamline digital analysis so agents can spend more time in the field?
Investigators were looking for a tool to help them analyze large overlapping email datasets that are received in multiple physical formats from various service providers.
Digital analysis is arduous
I was brought on board to address a pressing challenge in the field of investigations - the tedious and time-consuming process of parsing through thousands of emails in search of illicit content. However, through my engagement with field agents through research interviews and design thinking workshops, I was able to uncover a larger underlying problem.
This journey map followed the two agent personas who would use the platform.
Discovery
Ethnographic Research
By conducting ethnographic research, I was able to observe the tools that the law enforcement agents were familiar with and gain a deeper understanding of the nuances of their investigation process. One notable observation was the difficulties they faced in digitizing notes during nighttime investigation while hiding in vehicles, due to the absence of a dark mode option.
2 Day Champion Workshop
I co-facilitated a two day workshop in which the champion team discussed their experiences, worked in groups of two to map out the investigative journey, and express emotional and behavioral aspects of investigation through psychographic tokens.
Research Revelations
The ethnographic research and workshops facilitated a deeper understanding of the customers and their needs, allowing for a more effective solution to be created. I was able to identify that what investigators really needed was a tool that would allow them to document criminal lifestyle.
Clustering helped to guide feature brainstorms.
Designing for agents
Documenting criminal lifestyle
Collaboration across agents and operators
The champion team was specific that the tool needed to allow for operators to receive data in aggregate from ISPs, agents and other actors then analyze it and share the results with agents in the field. Similarly, agents should be able to collaborate with operators and add more data and information as cases are built.
Inspired by Microsoft Outlook
Watching the tools that agents use helped me understand what they were comfortable with. Most of their work was done using Microsoft Outlook on legacy Windows PCs. Other than that they used file managers and internal government systems.
At first I was horrified by the UI of Outlook and wanted to evolve it so that it would be elevated, but I soon came to realize that wasn't smart— I should design for my users not attempt to revolutionize digital data management.
During lunch conversations and design crits agents talked about what they wished they could do in Outlook. Core features included filtering, sorting, saving, tagging and commenting.
Filtering large data sets allows for a focused view of uploaded data.
Digital communications are organized
Process and Results
Brainstorming
As a team we did a brainstorming session where everyone will come up with a tool or solution related to one of the problems captured themes.
Expanded on the solution with two additional ideas based on personas designed.
Wireframes, feedback and iterations
Drew wireframes by hand and presented four potential solutions to the client. Then digitized and iterated on these wireframes, to account for existing familiar software.
After meeting with the development teams from Deloitte and Accenture I made some changes to the designs to prepare for handoff. After a few iternal design crits and final round of iterations, I presented my clickable prototype to the client.
Results
The client was thrilled with the final product as they could see how their agents' input had been incorporated. The ethnographic research and workshops facilitated a deeper understanding of the customers and their needs, allowing for a more effective solution to be created.
I was able to not only deliver the design deliverables expected of me, but earn the trust of my customers and help them ambassador the value of using design techniques to identify underlying problems. This led to an increased budget from the Department of the Treasury and contributed to their decision to keep Deloitte on retainer.
This user flow helped developers understand how documents were ingested, organized then analyzed in a familiar interface.
