Data Science and Human Trafficking


The United Nations (UN) has set “Decent Work and Economic Growth” among its 17 Sustainable Development Goals. Undoubtedly, the word “decent” does not correspond with forced labor, slavery, or human trafficking. According to the 2017 Global Estimate of Modern Slavery—developed by the International Labour Organization (ILO) and the International Organization for Migration (IOM)—25 million people around the world were victims of human trafficking for the purpose of forced labour in any given day of their lives. Although trafficking is a complex international crime that is very difficult to detect, science still has the upper hand and can offer humanity with potential means towards achieving its goals.

The Crisis

To clarify things for everyone, let us start with a comprehensive definition of human trafficking. According to the United Nations Office on Drugs and Crime (UNODC), trafficking in persons is the “recruitment, transportation, transfer, harbouring, or receipt of persons, by means of the threat or use of force or other forms of coercion, of abduction, of fraud, of deception, of the abuse of power, or of a position of vulnerability, or of the giving or receiving of payments or benefits to achieve the consent of a person having control over another person, for the purpose of exploitation”.

Human trafficking is a highly profitable organized crime that occurs all over the world, with profits second only to illicit drugs. Different countries around the world can serve as a source of victims, a transit point, or a destination and location of abuse. Trafficking victims are forced into slavery as prostitution workers, beggars, soldiers, factory workers, laborers in construction, mining, fishing, among others. Trafficking can be viewed as a supply chain where victims, representing “supply”, move through a network to meet the need for cheap labor force, representing “demand”.

However, these chains are dynamic, as traffickers tend to change their distribution and transportation routes to avoid detection, posing huge challenges to the law enforcement forces that try to stop them. The good news is that traffickers usually leave a data trail behind; whether bank transactions, photos on the Internet, online advertisements, phone calls, etc. Here is where science can interfere and use these pieces of information to break the supply chain at some point. Now, let us delve into more details about the different approaches used to combat trafficking in persons.

The Science

The science field involved in fighting this crime—Data Science—is relatively recent with a huge scale of promising applications in different fields; including marketing, scientific research, sports, agriculture, etc. In simple words, data science utilizes large amounts of complex data—big data—to draw meaningful information that assist in identifying regular patterns and making decisions. It is a multidisciplinary field that employs mathematics, statistics, computer science, artificial intelligence, machine learning, among others.

One approach of addressing human trafficking is helping banks trace traffickers’ money. For example, leading technology company IBM has recently developed a cloud-hosted data hub where banking institutions can provide enhanced information regarding suspicious money laundering activities intermingled with those from legitimate businesses. Using Augmented Intelligence (AI) and machine learning, the tool becomes more able to analyze the data and detect trafficking incidents. This consequently allows analysts to identify relevant information about them more easily. The effective sharing of such data would also allow governments and concerned institutions to take the needed action to shut down human trafficking.

A similar initiative has taken place in a collaboration between the Universiteit van Amsterdam (UvA), the Netherlands Ministry of Social Affairs and Employment, and ABN AMRO Bank to track the digital fingerprints of traffickers hidden in banking data, without violating the privacy of customers. The system developed at the UvA depends on 25 data indicators that help identify potential trafficking activity. For example, it tracks multiple account holders registered to the same address who make prompt withdrawals of minimum wages from the same ATM; those could mark potential victims. Bank data is analyzed every quarter to track these indicators, and the data is filtered regularly to secure privacy and prevent needless investigations. According to the UvA, the new project has thus far identified 72 suspicious bank account holders, and 50 potential victims.

Another approach of utilizing data science against trafficking is through preventing online communication between suspected traffickers and potential victims. In recent decades, vulnerable persons have been groomed through the Internet via fraudulent recruitment methods, such as promises of employment or marriage. Analyzing patterns of interaction on social networks could also help determine which contacts have a critical influence over others; this may enable early identification of traffickers and victims. In case the communication could not be blocked, data science still can help if any of the spotted potential victims went missing. It can provide investigators with the IP addresses and any available contact information of both the victim and predator, as well as other information regarding the communication between them.

Data science can model vulnerabilities of potential victims and help the concerned parties define the populations that could be approached by traffickers. In other words, it provides insight about which depressed areas or populations authorities need to target with anti-trafficking awareness campaigns and social service support. These vulnerability factors usually include poverty, unemployment, migration, and escape from political conflict or war.

A case in point is an initiative from India where poor village girls are usually targeted by traffickers with promises of better education, jobs, or marriage opportunities. Actually, the parents are not aware that their children are being sold into slavery; an Indian foundation known as “My Choices Foundation” has, thus, launched a program designed to raise awareness among villagers about how traffickers work. However, with over 600,000 villages in India, the program had to depend on a big data solution in order to identify the villages that are most at risk. The program developed by an Australian analytics firm analyzes India’s census data, government education data, and other sources for information about poverty level, proximity to police and transportation stations, etc., to draw the needed conclusions.

Another approach is to employ text analytics against trafficking. For instance, American analytics company SAS has supervised a project that utilized machine learning to assess patterns of human trafficking buried in the texts of hundreds of relevant official reports. The aim was to make these reports more accessible to the concerned organizations. Tom Sabo, of SAS, explains “We used text analytics to comb through all the Trafficking in Persons reports since 2013 and identify patterns that were not apparent previously”.

The findings included identifying the source and destination countries for trafficking worldwide across these years. Analyzers even drew color-coded lines between them to indicate whether a given pair of countries are cooperating to address the issue or not. Text analysis also involved spotting recurring word clusters to identify the purposes of trafficking in a given country. For example, a cluster of “forced”, “child”, “beg”, and “street” indicates exploitation of children as street beggars. The project provides researches and organizations with more easily accessible clues and figures from the reports, but they can still also resort to the supporting full texts to fully understand the context for the statistical findings.

The Challenge

Data science seems to offer a wide array of smart solutions, does it not? Yet, it still faces major challenges that hinder the urgently-needed fast action. One challenge lies in the fact that traffickers are so powerful in some countries that they threaten or bribe officials to ignore their vicious activities. As such, it is made easy for them to fabricate or alter the identification documents of their victims and make them invisible both to authorities and data analysts.

Second, but most importantly, data science needs “data” to function, and since the problem is global, the data needs to be global too. Unfortunately, it is still very hard to share data across different governmental agencies, NGOs, and data analytics institutions within the same country, let alone around the world. In addition to the sensitive nature of the information, there are no efficient and secure tools to facilitate sharing such huge amounts of data on a global level yet.

Trafficking is undoubtedly a horrible and horrifying crime; just like similar pressing global concerns, it requires serious actions and cooperation. It is high time humanity puts its trivial political conflicts and economic interests aside and work together towards a safer world where human beings are not treated as a commodity.


About Us

SCIplanet is a bilingual edutainment science magazine published by the Bibliotheca Alexandrina Planetarium Science Center and developed by the Cultural Outreach Publications Unit ...
Continue reading

Contact Us

P.O. Box 138, Chatby 21526, Alexandria, EGYPT
Tel.: +(203) 4839999
Ext.: 1737–1781

Become a member

© 2024 | Bibliotheca Alexandrina