There were an estimated 40.3 million people enslaved around the world as of 2016. That figure comes from The 2017 Global Estimates of Modern Slavery, the only high-quality global estimate of the prevalence of human trafficking. By contrast, there are 156,330 cases in the Counter-Trafficking Data Collaborative's global dataset, one of the largest sets of human trafficking data available.
Having a number of recorded cases which amounts to only 0.3% of the estimated prevalence seems like cause for despair, but used as data it represents a brighter future for counter-trafficking efforts. A number of innovators, ranging from data and computer scientists to national security analysts, have been creating incredible technology with the help of such data to assist in the fight against the multi-billion dollar slavery industry.
A.I. Classification
Contrary to the expectations of anti-encryption advocates in the 1980s and '90s, most child sexual abuse materials (CSAM) are shared in the open, on the clear web, with general purpose social media and file sharing platforms.
Survivor advocate Eliza Bleu has been working for the last year to bring awareness to the lax and inadequate response to CSAM on platforms like Instagram, Facebook, and especially Twitter. To get a grasp of how serious this problem is, over 60 million images and videos constituting CSAM were reported in 2020 alone.
Krunam
The above number comes from Krunam, a company that provides Artificial Intelligence for identifying and classifying CSAM. VigilAI, Krunam's A.I. powered image classifier, was trained on a Deep Learning model with data from the UK Child Abuse Image Database (CAID) and a carefully curated database of benign images.
Unlike older methods, the Krunam system is capable of classifying new images and videos, not just ones its seen before. It is also deployable via Docker, an application for compartmentalizing computer systems popular among cybersecurity professionals. By using VigilAI with Docker, an organization can easily deploy CSAM detection on a massive scale.
As well as classifying CSAM, VigilAI is designed and deployed to protect the psychological well-being of content moderators, as close contact with such materials can be psychologically stressful for them. Krunam also provides consultations on properly handling CSAM so as to balance the interests of the client business and the exploited person.
Artemis
The Global Emancipation Network's product, Artemis, is similar to Krunam's VigilAI in that it is an automated classifier. Artemis is distinct from VigilAI because it doesn't classify images, it classifies businesses. By searching for human trafficking keywords in reviews of common front-businesses, the system is able to give a probability ranking to each business, with a ranking of 1 indicating a 100% likelihood that the establishment is engaged in human trafficking.
Artemis is still in development, but the results of its pilot project are promising. The system looked at reviews for 3,567 Florida massage parlors, with 465 businesses being assigned a score 1. Even more impressively, the system identified 27 individuals operating massage parlors as part of Florida's human trafficking network. The G.E.N.'s ultimate goal is to extend the system to other industrial sectors, including social media content, to become a comprehensive system for identifying human trafficking operations.
Investigation and Intelligence
Minerva
G.E.N.'s other product, Minerva, is a robust investigation tool which is provided free-of-charge to qualified counter-trafficking investigators. Minerva has a wide range of investigative features, including advertisement analysis, language analysis, image processing, and a central database system.
By integrating tools from DarkOwl Cybersecurity, a company that specializes in dark web monitoring, the process of finding data from the deep web is made simple and secure for investigators who may not be familiar with this more complicated side of the internet. Coupling this with image analysis using Deep Vision's powerful A.I. architecture, public record searching, and geographical data allows investigators to spend less time doing analyses by hand. What's more, a central database structure used by multiple agencies is perfect for inter-organizational cooperation on human trafficking investigations
Hades
A similar product to Minerva is offered by the Anti-Human Trafficking Intelligence Initiative. Hades, their appropriately named dark web intelligence platform, is a means of hitting human traffickers where they'll feel it: in their wallets. Hades is geared towards cyber security businesses as a means of analyzing and stopping transactions related to CSAM.
Hades works by scraping dark web sites to find identifying information like IP and email addresses, encryption keys, keywords, metadata, cryptocurrency addresses, and other information that could help to identify people involved in human trafficking transactions. This gives cyber security experts the ability to perform financial investigations much more efficiently, without the need for complicated and expensive data collection and aggregation.
GOST
Giant Oak, Inc. is a risk management and intelligence company founded by national security experts. Their flagship product, GOST, is an AI intelligence platform for screening and vetting individuals geared towards businesses in general. GOST is designed to track an entity's online behaviour and aggregate multiple clear and deep web data sources into a unified view, customized for the user's industry. Using machine learning and behavioral science, the platform is able to detect high-risk individuals for client organizations to monitor.
In 2021 alone, GOST enabled its users to find 109 human and drug traffickers, along with hundreds of other fraudsters and violent criminals.
Traffic Jam
Traffic Jam, an investigation platform from Marinus Analytics, stands out from the rest due to its focus on victims of trafficking rather than traffickers. The most unique feature in the Traffic Jam platform is its ability to search vulnerability indicators.
Traffickers prey on the vulnerable, and trafficking is closely associated with social ostracism, poverty, and lack of adequate housing. Some of the most common tactics among traffickers are psychological manipulation and the restriction of movement. By enabling investigators to search based on vulnerability indicators, they can more accurately identify victims of trafficking and potential victims in high risk areas, such as those with many brothels and illicit massage parlors.
Traffic Jam also has a number of other highly advanced features, such as automated detection of potential trafficking and mapping of brothels and illicit massage parlors. One of the most impressive features of Traffic Jam is its facial recognition system, which allowed one investigator to put together a case in 3 months that would usually take 2 years.
Conclusion
Human trafficking is often poorly understood. It was as recently as 2000 that the Victims of Trafficking and Violence Protection Act was passed. Despite this, thanks to the efforts of technologists and activists, it seems there is a case for optimism. While human trafficking is still far from being abolished, advances in artificial intelligence and data collection have significantly bolstered the ability of counter-traffickers to fight back and legitimate businesses to protect their customers and their reputations.
As advances continue in the fields of artificial intelligence and more data on human trafficking is collected and put to use, it seems likely that counter-traffickers will have a significant lead in the technological arms race.