
With AI assistance, law enforcement could examine blockchain transactions from cryptocurrency wallets tied to digital trafficking.
Editor’s Note: The Red Cell series is published in collaboration with the Stimson Center. Drawing upon the legacy of the CIA’s Red Cell—established following the September 11 attacks to avoid similar analytic failures in the future—the project works to challenge assumptions, misperceptions, and groupthink with a view to encouraging alternative approaches to America’s foreign and national security policy challenges. For more information about the Stimson Center’s Red Cell Project, see here.
Red Cell
An adversarial relationship is developing between the technology industry and anti-human trafficking advocates. Human Rights Watch published an inflammatory blog declaring, “Using AI to Fight Trafficking Is Dangerous.” OpenDemocracy took aim at “Big Tech’s” efforts to create socially biased trafficking-detection tools that “[give] police more power [and do not] support trafficking survivors.” These anti-tech perspectives stem from widely confirmed data indicating social media and Web 3.0 technologies are causing a significant rise in human trafficking for sexual exploitation, organ theft, and forced labor.
Although they have good intentions, those who oppose the use of technology to combat human trafficking misdiagnose the condition. Technology is not the problem. The problem is the American strategy to employ it in this area and low expectations of the potential role of AI in combating human trafficking.
Understanding the Game
History’s darkest legacy is the trade of human beings. The seventh law of the Code of Hammurabi is the first written record of the slave trade. In the nearly 4,000 years since Hammurabi carved that law into stone, efforts to protect human rights in this area have not made much progress despite great efforts. The International Labour Organization estimates human trafficking generates $236 billion in illegal profits. Unlike other illegal trades like narcotics or weapons, forced labor generates recurring revenue at a very low cost to the seller.
Aside from the buyers and sellers, no one in human trafficking chooses to be a part of the trade. Most victims are tricked into joining with the promise of employment or safe migration. Some are kidnapped, but this is rare. Typically, human traffickers will (1) give the victim instructions on how to acquire a tourist visa, (2) bring them to the country where the trafficker operates, (3) confiscate the victim’s passport and electronics, and (4) move them to a new location. The trafficker effectively separates victims from their friends, family, and any means of escape. Violence, drug addiction, promises of vast riches, and coercion are commonly used to hold victims hostage.
Human traffickers view the trade as a game. They specifically target victims who are susceptible to exploitation. They build infrastructure, such as housing facilities and point-of-sale systems, to support their enterprise. Finally, they wager that governments will not detect the activity, under-resource counter-trafficking programs, and act slowly when a trafficker is exposed. Unfortunately, based on the strategy and investment priorities of the U.S. federal government, this is a safe bet.
Chips on the Wrong Numbers
Most of the U.S. government’s investment in anti-trafficking is largely aimed at reducing harm to victims. According to the Department of Justice (DOJ), 75 percent of the $361 million granted in fiscal year 2022 to counter human trafficking was allocated to victim services. No one disputes the value of victim treatment programs. Aside from addressing the moral imperative to assist people who have been harmed by trafficking, such programs help people escape exploitation, restore their dignity, and heal. These investments are rooted in decades of law enforcement experience with similar challenges, such as narcotics trafficking.
In the realm of illegal narcotics trafficking, many studies have concluded that reducing demand for illicit drugs is far more effective than incarcerating drug dealers. Treating consumers reduces demand, which removes the profit motive of the illicit drug trade and has historically generated positive outcomes. As the allocation of DOJ resources reflects, this logic has been applied to the human trafficking challenge. But the model simply does not fit. Unlike in narcotics trafficking, humans are the “product” in a trafficking enterprise—not the consumers. It would be absurd to send a kilo of cocaine (the product) into drug treatment. In effect, though, in the anti-human trafficking arena, that is precisely what is being funded.
The U.S. government is investing little to no effort or funds in reducing the demand for human trafficking. In fact, most federal and state laws are oriented toward punishing trafficking victims rather than the traffickers and customers. For those victims smuggled into the U.S. and forced into exploitative jobs, the U.S. fast-tracks their deportation. For U.S. citizens participating in sex work, policing is focused on prosecuting them rather than the traffickers’ customers. Congresswoman Ann Wagner (R-MO) introduced legislation to refocus federal law away from prosecuting trafficking victims and toward punishing perpetrators, but there has been no action on that bill since 2022. Some nongovernmental organizations (NGOs) are attempting to leverage technology to reduce demand, but they are a small minority in the overall human trafficking prevention ecosystem. This is a troubling trend because North America is one of the biggest markets for human exploitation.
Digitizing the Casino Floor
Human trafficking has several immutable features. Traffickers need to make money, and they need to advertise their business. In the past, traffickers thrived in the physical world. They financed their activities mostly with cash, drugs, or other tradable illegal commodities. Hand-to-hand, in-person transactions were standard. Illegal proceeds were laundered through cash businesses.
Cash is more expensive for businesses to manage than credit or debit transactions, and consumers use credit and debit cards more frequently. Cash-only businesses are anomalies and attractive targets for investigators. As a result, traffickers are moving toward electronic methods to capture transactions and launder money. Cryptocurrency and peer-to-peer payment systems are growing in popularity among traffickers because they sidestep financial institution surveillance systems.
Make no mistake: human traffickers view people as “products” in the digital economy. Perhaps the most repugnant example of this trend is the appearance of bar and QR code tattoos on the bodies of sex workers. These tattoos allow traffickers to track their victims, ensure customers provide valid payment before services are rendered, and conduct analyses on the performance of their business.
Human traffickers’ adaptation to the digital future is prominent on social media, job boards, and online dating platforms. Traffickers use these venues to recruit victims and advertise their “products” and “services.” These technologies also allow consumers to rate and review exploited persons.
An Unbalanced Roulette Wheel
The digitization of slavery presents a wide array of opportunities for law enforcement. Unfortunately, the U.S. government’s investment strategy to counter human trafficking is out of balance. Aside from the heavy emphasis on providing victim services, the U.S. government is not well-positioned to confront the problem.
The approach to trafficking investigations follows a traditional model. Most law enforcement agencies focus on individual traffickers, point-to-point transactions, and sting operations of known trafficking locations, such as massage parlors. Only a few law enforcement activities apply an “attack-the-network” approach to human trafficking. A few NGOs are helping law enforcement to use open sources more effectively, but their processes are largely manual and ill-suited to address the sheer volume of data.
Although the amount of online activity tied to human trafficking is unknown, the National Center for Missing and Exploited Children (NCMEC) has compiled some reliable data. The NCMEC operates the CyberTipline, which tracks suspected online child sexual exploitation. In 2023, NCMEC received over 36 million reports of suspected child sexual exploitation. In 2024, it received 27,800 reports of child sex trafficking.
This is just one subcategory of the overall human trafficking challenge. These numbers do not include adults trapped in prostitution, forced physical labor, or organ trafficking. These numbers also do not include all the digital advertisements, incidents of online elicitation, or blockchain transactions tied to human trafficking. With millions of reports at NCMEC alone, billions of pieces of valuable information are likely traversing cyberspace. Owing to the sheer volume of data, human-driven, manual investigative processes are doomed.
Machines Counting Cards
The answer is clear: the U.S. government should employ AI and automated data ingestion to assist law enforcement in prosecuting human traffickers. Since 2018, the NGO Tech Against Trafficking has maintained a database cataloging over 300 different technical tools to fight human trafficking. An analysis of this data indicates that just over 10 percent could be considered AI technologies. As of 2024, over half of those AI tools were no longer on the market because the companies were either out of business or no longer supported the platform. There are probably many reasons why this level of digital attrition is occurring. Without government investment in AI to combat trafficking, more attrition will occur, traffickers will invest more heavily in digitizing slavery, and there will be no hope of abating the scourge of modern slavery.
As noted above, activist groups object to the possible gendered, racial, and socioeconomic biases in some algorithmic platforms. Although these objections are valid, they fundamentally overlook how AI can be used for good. Activist groups are correct that most of the AI technologies designed to counter trafficking are unfairly aimed at victims. This is the same backward model of focusing on “the product” from the counternarcotics experience. Not only does this approach manifest the algorithmic bias concerns of anti-technology advocates but also it is ineffective because the victims often do not have a digital footprint. Traffickers exert control by turning their victims into “digital ghosts.” There is no value in using AI to focus on data that does not exist.
The more appropriate application of AI would be to focus on the activities, infrastructure, and digital dust of the traffickers. Borrowing a principle from the Defense Intelligence Agency’s SABLE SPEAR program, criminals want to be found. They just do not want to be found by law enforcement. The digital posture traffickers are adopting necessitates using phones and computers. Every digital element creates a cyber attack surface or intelligence-gathering opportunity. The challenge has been to investigate these activities comprehensively and find all the critical nodes of finance, transport, and housing in a timely manner. The current state of AI makes discovering these elements possible at machine speed and at a scale beyond human cognition.
With the assistance of AI, law enforcement could examine every blockchain transaction from every cryptocurrency wallet tied to a digital trafficking advertisement on the open or the dark web. Moreover, combining this mosaic with other contextual information in the advertisements would reveal specific locations and means of contacting the traffickers. Law enforcement officers could filter for information applicable to their jurisdictions and employ their investigative or interdiction authorities.
Finally, advances in generative AI could enable the creation of legal paperwork protecting civil liberties. For example, once a suspected criminal enterprise has been identified, search warrants, subpoenas, charging papers, and any other desired legal filing could be automatically created from digitally collected information. This would greatly reduce the administrative burden of bringing perpetrators to justice and accelerate the prosecution of suspects. This would also mean not “revictimizing” trafficked persons by forcing them to testify about their experiences in front of their assailants.
Despite the moral atrocity of human trafficking, criminal organizations view it as just one of their illegal businesses. The conveyances, routes, and methods follow similar patterns as weapons, narcotics, and wildlife trafficking. A pivot toward automated data collection and analytics is not a “nice to have.” It is vital to address this threat to human dignity and international security.
Brian Drake served as the Defense Intelligence Agency’s first Director of Artificial Intelligence and currently serves as Federal Chief Technology Officer of Accrete.AI Government.
The views expressed in this article are those of the author and do not necessarily represent the views of Accrete.AI. Neither the author nor Accrete.AI has any financial or business interest in publishing this article.
Image: Tinnakorn Jorruang / Shutterstock.com.