South Korea Enhances Missing Children Searches with AI

TLDRs;

• South Korea uses AI to generate adult images of long-missing children, aiming to boost search effectiveness.
• Super-resolution imaging improves poster clarity, helping authorities receive occasional new leads on cold cases.
• AI-generated projections have aided tips for children missing decades ago, though accuracy metrics remain uncertain.
• Collaborative government projects hint at growing opportunities for AI-based forensic technologies in law enforcement.

South Korean authorities are turning to artificial intelligence to tackle one of the most challenging issues in child safety: locating long-missing children. The National Center for the Rights of the Child (NCRC), a government-backed organization, has launched a project that uses AI to generate updated images of individuals who went missing years or even decades ago.

These AI-generated images, designed to show how a child might appear as an adult, aim to revitalize cold cases and generate new tips from the public.

The AI system, developed by the Korea Advanced Institute of Science and Technology (KAIST), leverages age-progression patterns to transform the last known photographs of missing children into realistic projections of their adult appearance. In one notable example, Kim I-gon, who disappeared in 1985 at the age of 13, now has a projected image depicting him at 52. Authorities hope these visual reconstructions will help reconnect families with long-lost loved ones.

Super-Resolution Imaging Sharpens Clues

KAIST has enhanced its age-progression technology with super-resolution imaging, a technique that improves image clarity and detail. This upgrade allows generated posters to capture subtle facial features and expressions, making them more recognizable to the public.

“While these AI-generated images cannot guarantee identification, they keep missing children cases visible and can spark new leads,” said an official involved in the project. To date, the NCRC has produced AI posters for 60 long-term missing children, and some have already led to tips from the community. Similar AI-based efforts have recently been implemented in countries like Argentina, indicating a growing global interest in technology-assisted child recovery.

Accuracy Remains an Open Question

Despite the potential, experts caution that the accuracy of AI age-progression remains difficult to measure. The (NCMEC) in the U.S. reports that its forensic artists completed over 7,800 age-progressed images, helping recover more than 1,800 children. However, official validation studies and accuracy rates are limited. Traditional methods, which rely on skull development patterns and family photographs, continue to play an important role in generating realistic projections.

“The technology is evolving, but it’s important to understand it as a tool for generating leads rather than providing definitive identification,” said a child welfare specialist. The posters primarily function to maintain public awareness of unsolved cases and encourage tips from potential witnesses.

Government Collaboration Spurs Tech Opportunities

The South Korean initiative is a collaboration among the NCRC, the National Police Agency, and the Ministry of Health and Welfare. Experts note that projects like this highlight potential opportunities for AI forensics vendors, as governments seek advanced imaging and investigative tools to support law enforcement.

From age-progression systems to super-resolution imaging and public campaigns such as “Runway to Home,” these multi-capability platforms could become standard in child recovery efforts. The move underscores how AI is increasingly applied in public safety contexts, bridging technology with traditional investigative methods.

As South Korea continues refining its approach, authorities remain optimistic that AI will help bring more long-missing children back into the public eye and, ultimately, to their families.

The post South Korea Enhances Missing Children Searches with AI appeared first on CoinCentral.