How a 20-year manhunt shows the rise of tech-driven anti-trafficking in China

On March 21, 2026, Chinese authorities announced the arrest of a suspect known as "Mei Yi," bringing an end to one of the country's longest-running child trafficking investigations.

The case spanned more than two decades. It began in 2003, when a series of child abductions first surfaced. Over the years, victims were gradually identified and reunited with their families, with the last case as recently as 2024. What remained unresolved was the identity and whereabouts of the alleged trafficker at the center of it all.

By the time investigators closed in, "Mei Yi" no longer closely resembled the figure in early sketches.

According to people familiar with the case, her appearance had changed so significantly over two decades that it bore less than 30 percent similarity to the widely circulated composite portrait.

That gap between who she was and how she appeared is precisely where new technologies began to matter.

In the "Mei Yi" case, authorities relied on a combination of DNA databases, cross-age facial recognition powered by artificial intelligence, and large-scale data analysis to narrow down leads. The ability to match childhood images with adult faces proved particularly critical.

Over the past decade, China has expanded its national DNA database to reconnect families separated by trafficking, while AI tools are being trained to recognize individuals across decades of physical change. At the same time, big data systems are used to map movement patterns and identify potential networks, reducing reliance on fragmented clues.

China's anti-trafficking tech

For human trafficking cases, investigators usually first turn to facial recognition systems to scan surveillance footage across high-traffic locations, including railway stations, airports, bus terminals, as well as shopping malls, schools and hospitals. Some transport hubs are equipped with systems capable of processing thousands of faces per second, automatically flagging individuals whose features match known trafficking suspects. Images of missing children are also continuously compared against live feeds.

But identifying a match is only the beginning.

The technical challenge lies in the fragmented nature of visual data. Footage from different sources varies widely in quality, angle and lighting. Once potential leads are generated, they must be cross-checked against existing suspect databases. This stage still relies heavily on human investigators, as matching accuracy remains constrained by algorithm performance and computing power.

In recent years, additional tools have been introduced. Gait recognition, which identifies individuals based on body movement patterns, has been deployed in criminal investigations to supplement facial analysis when images are unclear or incomplete.

More advanced still is cross-age facial recognition. By simulating how a face evolves over time, AI models can estimate what a suspect might look like decades later. However, challenges still remain, particularly around genetic variability and the unpredictability of aging.

Ultimately, the most definitive tool is DNA matching. In China, a national DNA database established in 2009 by public security authorities to locate trafficked and missing children has played a critical role in reconnecting families separated for years. Through matches in this database, thousands of children abducted years ago have been identified and reunited with their biological parents.

While human trafficking remains a global challenge, China's approach – combining centralized databases, AI tools and large-scale coordination – reflects how technology is being used to address it.

For families involved in the "Mei Yi" case, however, the significance is more immediate.

After more than 20 years, the search has come to an end.