Swedish Lab's AI Spots Concealed Guns Through Wi-Fi Signals
Researchers at the Swedish Defense Research Agency (FOI) built an AI system that identifies concealed handguns by reading ordinary Wi-Fi signals. The system, called WiFireNet, needs no camera line of sight and no walk-through metal detector. The team described the design in a paper posted to arXiv on October 13, 2026. The setup has three parts: a Wi-Fi transmitter, a Wi-Fi receiver, and the WiFireNet model. The model analyzes how objects distort the signal passing through a space. It distinguishes a plastic gun from a metal one, and both from other everyday objects. FOI pitches the work as a non-intrusive alternative to visible cameras and metal detectors. Carry-rights advocates should read that pitch twice.
Key Details
The published work contains the following specifics:
- Developer: FOI, Sweden's defense research agency.
- Publication: A paper posted to arXiv on October 13, 2026.
- System name: WiFireNet.
- Hardware: A Wi-Fi transmitter, a Wi-Fi receiver, and the AI model.
- Input data: Channel state information, the measurable way an object alters ambient Wi-Fi as the signal reflects off it or passes through it. Shape, orientation, and material all leave a signature.
- Targets: Concealed handguns, with the model trained to separate plastic and metal firearms from other objects.
- Training method: Paired video and Wi-Fi signal recordings taught the model what each object looks like in signal form.
- Related prior work: Security researchers previously showed that ordinary Wi-Fi routers can identify people with 99.5 percent accuracy.
The reported results focus on detection, not enforcement. Nothing in the coverage describes a deployment, a government contract, or a commercial product. The concept works in a lab. Whether it works in a crowded lobby, through a wall, or on a moving target remains unproven in the material available. The 99.5 percent figure comes from a different research effort on person identification, not from WiFireNet itself.
Background
Wi-Fi sensing is not new. Researchers have used channel state information for years to detect human presence, track movement, and recognize gestures. Elder-care and home-monitoring products have leaned on the same physics. WiFireNet applies that research to a narrower target: the metal and polymer mass of a firearm under a shirt.
Automated surveillance has a recent record that gun owners already know. Flock Safety's license plate reader network drew heavy scrutiny from civil liberties groups over data retention and sharing with law enforcement agencies. Many of those cameras sit in private hands, including businesses and homeowner associations. That pattern of private collection feeding public databases is the precedent that matters here.
The legal backdrop comes from New York State Rifle & Pistol Association v. Bruen (2022). The Supreme Court struck down New York's proper-cause requirement for public carry. It established that modern gun laws must be consistent with the nation's historical tradition of firearm regulation. The majority also said some sensitive places could be restricted, while leaving the list open-ended. Most states already let private property owners ban firearms on their premises by default or by posted notice.
Why It Matters for Gun Owners
Concealed carriers face the most direct exposure. A detection system that flags a holstered handgun turns every carrier into a flagged object, whether or not they broke any law. Carriers in states with broad sensitive-place rules, and those in states where private owners can ban carry by default, should assume that detection tech will spread into those venues first. Schools, hospitals, stadiums, bars, and transit hubs are the obvious early buyers.
Private businesses will not need a new law to act on this. A store owner with a posted no-carry sign can ask a customer to leave, and a detection system gives that owner a reason to check. Expect false positives. Water bottles, laptops, tools, and phones all change signal patterns. A system that cannot tell a gun from a hard case will generate confrontations.
Watch three things. First, procurement: look for city or county contracts for Wi-Fi sensing, particularly where municipalities offer public Wi-Fi. Second, data retention: ask whether any detection logs get stored or linked to identity. Third, state legislation on biometric and sensor-based surveillance, which could cover this technology if drafted with care.
DownRange Analysis
A detection algorithm is not a gun law. Bruen asks whether a regulation fits the historical tradition of firearm restriction. WiFireNet regulates nothing. The Fourth Amendment question is where this gets serious. In Kyllo v. United States (2001), the Supreme Court held that using thermal imaging to peer into a home without a warrant counts as a search. A government agency using through-wall Wi-Fi sensing to find guns inside a home would press that same line. Gun owners should push courts to extend Kyllo to this tech before agencies deploy it.
The market will follow the pitch. Sensor vendors will sell gun detection to venues that already pay for metal detectors and cameras. Cheaper, invisible sensing will look attractive to budget-strapped districts. Expect that sales pitch long before any court weighs in.
What should a gun owner do right now? Keep carrying under your state's rules. Know your state's default private-property rule and read the signs at every door. Do not treat a lab paper as a shipped product. Contact your state legislators and ask them to address government use of sensing technology before it arrives. Track procurement records in your city. The fight over this tech starts in budget meetings, not courtrooms.
