Won’t somebody please think of the children?! Well, there’s no need this time, because it seems the children are already thinking of us. Hsu Ting-wei, a 10-year-old from Taiwan, has just won a gold medal and the Best Invention Award from the International Federation of Inventors’ Associations (IFIA) for a device that looks like something out of a movie: a sort of “electronic nose” that analyzes the air around luggage and, according to its creator, could one day help detect drugs like fentanyl.
So, did a kid invent an AI that can sniff out drugs hidden inside a suitcase? Well… not exactly. But the real story is hardly less interesting.
Hsu, a student at Kang Chiao International School in Taiwan, presented his “Novel Artificial Electronic Nose” at the 2026 Silicon Valley International Inventions Festival in Santa Clara, California. The judges not only awarded him a gold medal, but his project also won the IFIA Best Invention Award.
Yes, amazing: a 10-year-old developing a fairly advanced detection device—perhaps even comparable to systems currently used at some airports—in the name of science. Amin Ghafooripour, the head of the jury, said Hsu received the recognition because the judges saw practical value in his proposal.
Electronic noses, AI and drugs: How does the invention work?
The device works, in a sense, like an artificial nose: it samples the air surrounding a piece of luggage and looks for chemical information that could indicate what might be inside.
The system is intended to serve as a follow-up step after airport X-ray screening and similar security checks. Tubes draw in the air around the suitcases, and the sample is analyzed using gas chromatography-mass spectrometry (GC-MS), a technique widely used in laboratories to separate and identify chemical compounds.
The system also incorporates AI-based recognition capable of learning patterns associated with specific substances. Rather than replacing X-rays, the idea is to add a second layer of information: first looking inside a suitcase and then—figuratively speaking—”smelling” it.
But here’s the small detail some headlines overlooked: drugs were not what the invention was originally designed to detect.
Hsu had already presented a version of his project at the 2025 International Exhibition for Young Inventors under a somewhat less spectacular name: Novel Artificial Electronic-Nose: Pork Scent Detector. Yes: a machine designed to smell pork.
The goal was to identify pork products hidden in luggage—something particularly relevant for countries trying to prevent the entry of food products that could spread diseases such as African swine fever. In that demonstration, the system scanned for chemical compounds associated with the scent of pork and routed suspicious luggage to an inspection area.
The project won gold then, too. As early as 2025, Hsu suggested that, when combined with AI, the same concept could be extended to drug detection.
So, can it smell fentanyl?
This is where we need to dial the expectations back a bit. In California, Hsu said he hoped his system could help the United States intercept substances like fentanyl at its borders. Project descriptions also mention the possibility of training the AI to recognize new chemical profiles associated with drugs.
But that is not the same as proving there is currently a machine capable of sampling the air around any random suitcase and shouting, “Fentanyl!”
At least in the publicly available information about the project, there is no mention of fentanyl detection trials under real-world conditions, accuracy rates, detection limits, or data on false positives or negatives. The documented 2025 demonstration involved pork products.
In other words, while the architecture could potentially be applied to drug detection, whether the system can already detect drugs reliably is another matter.
In any case, techniques for analyzing minute chemical traces already exist, and the concept of using chemical data to complement other security systems is not new. What is novel about Hsu’s project is the way it brings together automated air sampling, GC-MS, computational analysis, and an inspection workflow designed specifically for luggage.
Other electronic noses that are already learning to smell
Hsu isn’t the only one trying to digitally replicate one of our most elusive senses.
So-called “electronic noses” (or e-noses) have existed for decades and aim to mimic the olfactory system by detecting volatile substances and turning their signals into patterns a computer can recognize. Their applications range from food and environmental pollution to medicine and perfumery.
In 2025, for instance, researchers from the Daegu Gyeongbuk Institute of Science and Technology (DGIST) in South Korea published a study in ACS Nano describing an electronic nose that uses tiny sensors made of graphene and cerium oxide. Different odors generate distinct combinations of electrical signals, and a machine learning model learns to distinguish between them. In controlled tests, the system identified nine odorant molecules used in perfumes and cosmetics with over 95% accuracy.
However, this isn’t exactly the same thing Hsu invented. The Korean project focuses on developing new sensors designed to behave like olfactory receptors. Hsu’s device relies on GC-MS to chemically analyze the air and then incorporates AI to interpret that data within a screening system.
So no, there probably isn’t a 10-year-old roaming airports just yet with an AI capable of sniffing out suspicious little packages. But there is one trying to teach a machine to decide which suitcases warrant a second look. And, considering he started out wanting to detect hidden pork and is already thinking about fentanyl, his work might be worth keeping an eye on.
Cover photo created with AI.
