New AI Drug Test Needs Just One Drop of Urine and No Prep

South Korean researchers built a gold nanopillar and AI platform that screens a single drop of urine for illicit drugs in five minutes.

AI Tools
New AI Drug Test Needs Just One Drop of Urine and No Prep

A team of South Korean researchers has built a testing platform that can screen a single drop of urine for illicit drugs in about five minutes. The system pairs a gold nanostructured surface with an artificial intelligence model. It needs no chemical pretreatment of the sample.

The work was led by Dr. Eun-Kyung Lim at the Bionanotechnology Research Center of the Korea Research Institute of Bioscience and Biotechnology (KRIBB). Partners included Professor Tae Hwan Kim of Daegu Catholic University, Professor Joon Hee Lee of Seoul National University Bundang Hospital, and the AI Research Center at StoneLab.

The study was published in September 2026 in the journal ACS Measurement Science Au. Dr. Jinyoung Kim of KRIBB is a first author.

The Gap Between Lab Tests and Quick Kits

Drug testing today usually falls into two groups. Laboratory tests are accurate, but they need several sample preparation steps and specialized equipment.

Rapid on-site kits give faster answers, but their accuracy has been limited. Problems include small sample volumes and cross-reactivity, where a kit reacts to drugs with similar chemical structures.

The researchers say their platform brings together strengths of both approaches. It aims to pair lab-style precision with the speed of a test kit.


How Gold and AI Find the Drug Signal

The core of the system is a thin gold surface covered in densely packed nanopillars that look like blades of grass. The team roughened the pillars through a process called selective etching.

When drug molecules land on this surface, the faint light signals they give off are amplified by millions of times or more. This lets the system pick up the vibrational “fingerprint” of a drug, even in trace amounts.

Urine is a hard sample to read because it holds many metabolic waste products and traces of medications. To sort through that background, the team used a machine learning model called LightGBM. The model was trained to tell the signal patterns of illicit drugs apart from the other signals in urine.

Coverage of the research has noted that experts see the proof of concept as promising. They have also said more validation will be needed across a wider range of drug classes and patient groups.

In tests with real urine samples, the platform identified drug-positive samples with about 80% accuracy. That result held even for samples from patients taking non-narcotic painkillers or several medications. In rat experiments, drug components could be analyzed in urine collected within six hours after the drugs were given.

The team says the next step is turning the platform into a compact, portable device. Possible uses include on-site screening by law enforcement and fast checks for drug intoxication in hospital emergency rooms.

Further work is expected to test the platform on more compounds and clinical samples. Plans also include scaling up chip manufacturing and making the AI model’s decisions easier to interpret.

The research was funded by several South Korean agencies, including the National Research Foundation of Korea and the Korea Medical Device Development Fund. KRIBB’s major research programs also provided support.

“This study demonstrates that combining nanoengineering with explainable AI algorithms can enable rapid, precise classification of illicit drugs even in complex bodily fluids,” Lim said. She added that the team hopes to “transform drug screening and emergency diagnostics” by developing a portable version for on-site use.

From our research desk
AI Jobs Automation Index
Which jobs are AI tools targeting most? We mapped 3,400+ AI tools to real job functions — with BLS employment & salary data.
Explore the index
maisiekooc

Written by

Maisie is a news writer at Agent Locker, covering the latest developments in artificial intelligence, emerging technology and the companies shaping the future.

Discover AI Agents