Skip to content
Forced labor/Supply chain

Updating human rights law necessary to combat “digital forced labor” in age of GenAI

Ann-Elise Quinn  Analyst / Thomson Reuters Special Services

· 5 minute read

Ann-Elise Quinn  Analyst / Thomson Reuters Special Services

· 5 minute read

The United States must adapt its laws and law enforcement strategies to prevent the use of data derived from forced labor in artificial intelligence systems and emerging technologies

While the United States is presently a global leader in irradicating goods produced by forced labor from entering supply chains, our current import laws may be insufficient to prevent data derived from forced labor from being built into our artificial intelligence (AI) systems, generative AI (GenAI), and other emerging technologies.

Indeed, in the age of GenAI, the US must adapt its laws and law enforcement strategies to continue to be a strong advocate for global human rights.

Many technological optimists believe that GenAI will bring solutions to some of humanity’s most challenging problems. They believe that AI may eventually be capable of collecting disparate bits of information from all over the world to quickly recognize patterns and make inferences that human minds may not. These tools may augment our abilities in medicine, science, agriculture, and innumerable other areas.

Touting the positive outcomes of AI in human rights

For those who are interested in promoting human rights and fair labor practices, there is hope that AI may be a tool to promote fair labor practices. For instance, an AI system may be trained to identify supply chains, regions, and particular factories that are likely abusing their workers and produce goods through the use of forced labor. A tool like this would be invaluable to US companies and consumers who do not wish to purchase goods made by forced labor. Indeed, not only is it illegal to import these products into the US, but also most people (and companies) with a conscious would not wish to contribute to this human rights crime.

Presently, the inherent complexity of global supply chains and the fact that the victims of these practices are often those with the least ability to speak out can cause these crimes to be difficult to be identified. A supplier may perpetrate these crimes for years before its importers are aware. However, AI could quickly recognize indicators of forced labor within a supply chain and alert customers and relevant authorities.

Tech pessimists see new exploitation potential

Sadly, with every new technology, people have found new ways to exploit others. AI has already been used to perpetrate labor abuses upon vulnerable people. In 2022, researchers from the Distributed AI Research Institute identified instances of labor abuse for those that work as data labelers in the development of the large language models necessary to the development of AI systems. Data labelers often are paid low wages to perform repetitive labeling tasks, and much of this work is performed in regions outside the US. Indeed, labor practices or minimum wage laws may be of little use in protecting these workers, especially if they are in countries without robust labor laws and are ripe for labor exploitation.

When dealing with tangible products, the US bans the import of these goods and thereby creates economic consequences for these labor abuses. Often countries and companies will seek to rectify these practices so that they might continue trade with the US. When dealing with intangible products such as data, however, it becomes less clear if the US can or would prevent the use of data derived from forced labor within a US- based AI system.

Trade laws such as Section 307 of the 1930 Tariff Act and the more recent Uyghur Forced Labor Prevention Act are meant to prevent products coming into US ports that have been “mined, produced, or manufactured wholly or in part by forced labor.” But when the product — the data — can be transported essentially through an email, the policies and enforcement actions currently in place may fall short. These laws were originally meant to govern tangible goods rather than intangible goods that don’t require traditional ports of entry.

Perhaps these laws could be interpreted to infer the inclusion of intangible data; however, a clear law stating that data derived from forced labor is not acceptable in the US would clarify any question on this topic and further establish the nation’s commitment to combatting forced labor.

To keep pace with technology and its commitment to human rights, the US should institute laws and policies that make it clear that any data derived from forced labor or labor abuses may not be used within the country for any purpose. Many critics would rightly state that this would be a challenging rule to enforce. However, policies like this would be in the spirit of the laws already in effect and would clarify the nation’s commitment to labor rights in the coming age of AI and GenAI.

Setting up consequences for bad behavior is among the major purposes for laws regardless of whether that bad behavior is easy to identify. Practical steps for enforcement will come when policymakers have provided a need.


You can find more information about ways to combat forced labor within organizations’ supply chains here.

More insights