The Need for Artificial Intelligence in Human Trafficking Research
The applications of AI have been widespread in the last few years. We have seen unique applications of AI in biology, sports, math, science, art, education and even music. AI is impacting self driving cars and even predicting recidivism rates amongst convicts… but how about human trafficking?
What is human trafficking?
Before we know how we can make an algorithm for human trafficking, we need to know exactly what it is. According to the Department of Homeland Security, human trafficking involves the use of force, fraud, or coercion to obtain some type of labor or commercial sex act. Contrary to popular belief, it does not always involve moving people across countries or even state borders. Human trafficking can take place in your own home. Oftentimes friends, partners, and family perpetrate this crime.
As per the definition of human trafficking, there are two general categories: sex trafficking and labor trafficking. These nuances are important to consider when building algorithms for human trafficking.
Why is there a need for artificial intelligence in human trafficking?
Human trafficking is hard to prosecute
Human trafficking is one of the least prosecuted crimes in the world because of the complexities of the crime. In the case of labor trafficking, it is hard to draw the line between poor labor conditions and forced labor. In the case of sex trafficking, survivors often feel shame or guilt regarding their situation, and are hesitant to speak out. Sex trafficking acts are also often masked as businesses such as massage parlors and spas.
Survivors also often find that they would not gain anything by speaking out. Traffickers often coerce individuals into believing that they have a debt which they can pay off through working under the trafficker, or promise some sort of future incentive to keep working such as a green card. In any case, traffickers incentivize their victims – no matter how small – and that keeps victims compliant. In seeking justice, given the fact that trafficking is hard to prosecute, their perpetrator may not even be found guilty. Instead, the survivor may face further abuse from the perpetrator as a result of speaking out.
There is a lack of human trafficking data available
As can be expected, there is not enough data publicly available about human trafficking. The trafficking numbers around the world is reported to be low, and according to some sources, the numbers are in the single digits, despite other data showing large trafficking networks in those countries. Moreover, many trafficking reports fail to show the type of trafficking taking place. In fact, it is not uncommon to see data tables with missing columns and rows.
AI can help fill on those gaps
In light of the many applications of AI that deal with social and criminal justice issues, it is not unbelievable that AI could help human trafficking issues. Given that there are more qualitative reports about trafficking available on the internet as opposed to quantitative, with the use of natural language processing, you can predict the presence of trafficking. Furthermore, in using algorithms like tf-idf, bag of words, BERT and convolutional neural networks (CNN), you can analyze trafficking reports and articles to find new meaning. As an example, in this blog, I detail a labor trafficking detection tool I made with Inspirit AI. The tool is trained on reports and articles and uses CNNs to predict the presence of trafficking. AI can also be used to map geographic patterns that identify trafficking victims or perpetrators on the move. It can also be used to make heat maps indicating where trafficking is occurring the most – and these applications are only the beginning of what’s possible.
Who should get involved with artificial intelligence in human trafficking research?
Anyone can get involved! If you are someone who is more interested in social science, looking at human trafficking applications of AI can help you see the world through a more statistical lens. You can grow skills in data analysis and experimental design, which will be immensely helpful if you see yourself working at non profits down the line. If you are a programmer or prospective data scientist, human trafficking research can help you build skills in machine learning that you were previously not familiar with such as natural language processing. Someone who is already involved in social work even as a human trafficking specialist can expose themselves to a new way of finding information about human trafficking through AI.
If you are a teenager or younger adult who is interested in this application of AI, you can consider working with someone who has a social work license. Human trafficking is a heavy topic, and you want to make sure you are not emotionally harming yourself in the process. They can help you make sure you are emotionally safe during the duration of your research or project. If you are planning to conduct interviews or work with human subjects, you may need a certification or license. Make sure to check the requirements, as they may be very subjective based on the exact application of AI you choose.
Examples of using artificial intelligence in human trafficking research
There are a growing number of ways that you can apply AI to human trafficking research, whether it be through natural language processing, computer vision, or something else! Here are some you can consider:
A logistic regression classifier that predicts whether trafficking is taking place or not
A logistic regression classifier that predicts how likely human trafficking convicts are to recidivate.
An algorithm that analyzes images of labor sites to determine if they indicate trafficking
An algorithm that analyzes images of sex advertisements to determine if they indicate sex trafficking
An algorithm that predicts trafficking risk based on location and identity (socioeconomic status, gender, race, etc.)
Organizations that tackle human trafficking
As you think of your next steps in human trafficking research, you may want to look into working with organizations that are working in prevention or simply read more about the issue. Here are some organizations that specialize in human trafficking research with AI, or work with human trafficking as a whole.
Organizations tackling human trafficking with AI
Marinus Analytics - Traffic Jam
Marinus Analytics builds technology tools geared towards law enforcement to fight human trafficking. They are partnered with more than 80 law enforcement agencies, and built Traffic Jam to find missing persons and stop trafficking. Some of the most impressive features of Traffic Jam include their geo-location mapping feature which maps brothels and massage parlors, which are areas where trafficking is most prevalent.
Street Grace - Gracie AI
Gracie AI is a trafficking prevention tool that won the Shorty Award in artificial intelligence. It combats the issue of online child sex trafficking in particular. Gracie is an artificially intelligent “human” which impersonates minors selling sex to lure predators to chat with them. Gracie AI then collects data about them and sends it to law enforcement.
Organizations tackling human trafficking in North America
Polaris
Coalition to Abolish Slavery and Trafficking (CAST)
DeliverFund
Organizations tackling human trafficking globally
Hope for Justice
Stop the Traffik
International Justice Mission
Free the Slaves
A21
The Exodus Road
Destiny Rescue
Hagar
Unseen
Your next steps in experimenting with artificial intelligence in human trafficking research
Some of the best ways to understand artificial intelligence in human trafficking is to explore what is already being done by looking into organizations and research. In doing so, you can pinpoint some of the areas of trafficking prevention that still need improvement, and try to get involved in your own projects or research to alleviate the issue. You could even partner with, or join an organization to better their work. Whatever you choose to do, know that there are numerous ways that you can use your skills to make a positive contribution.
About Inspirit AI
AI Scholars Live Online is a 10 session (25-hour) program that exposes high school students to fundamental AI concepts and guides them to build a socially impactful project. Taught by our team of graduate students from Stanford, MIT, and more, students receive a personalized learning experience in small groups with a student-teacher ratio of 5:1.
