J{AI}NE DOE

Collaborative start-up project around online human trafficking. Selected for Imperial Venture Catalyst Challenge 2021

Date
May 2021
Role
Conceptualisation
Animation
Communication
Strategy
Project Type
UX Design
Strategy
Service Design
Featured
Reuters
[Aug 2021]
Overview

J{AI}NE DOE, a digital tool to extend law enforcement’s ability to reach, identify and convict online sex traffickers by streamlining, automating, and scaling the deployment of AI victim personas to engage online sex traffickers.

Problem Space

The internet is a hospitable environment for traffickers as they can easily scale their recruiting and grooming operations, law enforcement presence is minimal, and anonymity helps them avoid detection while abetting the deception of potential victims.

Approach

J{AI}NE DOE seeks to make the online environment more hostile for traffickers and serve as deterrence by supporting cyber operations and swarming the online platforms with AI victim personas.

Selected for:

Featured in:

Project Outcome

A digital platform that allows law enforcement to augment their ability to reach, identify and convict traffickers. It streamlines, automates, and scales the deployment of AI victim personas to engage with online sex traffickers. At the moment, law enforcement is outnumbered and out-gunned on the internet, but J{AI}NE DOE aims to change that.

How it Works

J{AI}NE DOE assists law enforcement with three key stages of running cyber stings: Generate, Engage, and Monitor

Generate

Cyber operators can generate and embed fake personas of potential victims based on specific parameters such as location, characteristics, etc. Or by using templates before further customizing any details.

Engage

Once deployed, the personas can support operators and/or engage directly in conversations with the individuals reaching out to these fake victim profiles.

Monitor

AI models trained on historical trafficker-victim conversations will flag interactions that fit trafficker behavior for further analysis.

The behaviours system supports analysis by generating summaries of key interactions that indicate potential trafficking behaviours such as recruiting and grooming.

Process

Understanding the problem

We have to question why traffickers feel so bold online when in real life, they may hesitate to approach a child on a sidewalk knowing they risk being noticed by the community around them. It’s no surprise that the anonymity of the internet makes recruiting and grooming online far safer. And this bears out in the statistics as 55% of victims trafficked in 2015 never even meet their trafficker.

Pain points

J{AI}NE DOE targets common pain points in running cyber sting operations

from helping users sound and behave like potential victims, each with their own unique personas, to supporting the social media content creation process.

This includes content scheduling as detectives will often spend their nights and weekends posting since convincing decoys don’t clock out at 5 pm.

Competitive Analysis

Industry Engagement

Our team conducted extensive research on stakeholders, interviewing former police chiefs, advocates for children and survivors, to machine learning experts. In the process of doing so, we’ve found that our different approach has really resonated with these domain experts, which has allowed us to gather a well-rounded advisory board. 

Stakeholders Insights

Designing the platform

A cyber sting social media management tool is a great launchpad for our grander vision. By tightening the scope of our MVP to offer something as simple as possible yet helpful, we get our foot in the door early and begin collecting the data needed to train machine learning models that will improve J{AI}NE DOE’s capabilities over time.