Between April 22 and 29, 2026, staff at Collins Bay Institution in Kingston seized multiple packages of drugs, tobacco, weapons and cellphones. CSC said the packages are believed to have come in by drone drop, and put their institutional value at about $171,806. It was the fifteenth contraband seizure CSC had reported at Collins Bay that year. Thirteen months earlier, Ottawa had quietly given the Correctional Service of Canada something Quebec’s public security minister had publicly called urgent: the legal authority to jam drones. Whether that authority has stopped a single drop, nobody outside government can say.

That is the real story, and it is not the one most people in the drone community believe. The common line is that nobody in Canada is legally allowed to stop a contraband drone. That stopped being true in March 2025. What replaced it is a more uncomfortable problem: a temporary pilot, limited to two correctional services, with no public accounting of whether it works — built around a tool that the next generation of drones is designed to ignore.

The Authority Already Exists

On March 3, 2025, the Minister of Industry registered the Radiocommunication Act Exemption Order (Jammers — Certain Correctional Services), SOR/2025-63. It exempts employees of CSC, and of Quebec’s provincial correctional services, from the Act’s prohibitions on installing and using jammers. The purposes are spelled out: public safety at penitentiaries and prisons, investigation of offences, and prevention of serious harm.

The Regulatory Impact Analysis Statement is explicit about drones. Continuous jamming is intended to block contraband cellphones inside institutions, while sporadic use can counter drone activity, with emissions directed beyond the perimeter where needed. CSC must notify NAV CANADA at least 28 days before first use at any site, restrict interference to the smallest area and shortest duration practical, and log every activation.

The RCMP has held a jammer exemption since 2015 and DND since 2024, but SOR/2025-63 was the first granted to correctional services, and the first extended to a provincial organization.

Why Ottawa Moved

The numbers CSC submitted to justify the order are stark. By CSC’s own count, drone-related incidents at federal institutions rose from 266 in 2020–21 to 619 in 2022–23. Cellphone seizures nearly tripled over a similar window, from 305 to 846. In late 2024, the CSC Commissioner said all federal institutions had drone detection. Detection showed staff the drones. It did not stop them.

The problem has not eased since. Joel Blacklock, a CSC senior project manager for counter-drone technology, told CBC in October 2025 that incidents reached 899 in 2023–24 and 1,064 the following year. At Collins Bay alone, CSC reported seizures worth about $164,375 in early February 2026, $158,090 in late February and $85,460 in early March. The Walrus reported in August 2026 that CSC has dealt with almost 3,000 drone incursions nationally since 2023.

Why Jamming the Link May Not Stop the Drop

Clarion has flown BVLOS operations outside Canada since 2018. One lesson from that work applies directly here: a drone that loses its control link does not fall out of the sky. It does what it was programmed to do.

Jamming the command-and-control link triggers the aircraft’s lost-link failsafe. Depending on the platform and how the operator set it up, that means return-to-home, hover, land, or continue the mission. Some flight software lets a waypoint mission carry on after signal loss. A contraband drone flying a pre-planned GPS route to a drop point does not need the link once it is airborne. Cut the link, and the package may still land in the yard.

Stopping an autonomous drone electronically generally means denying it satellite navigation through GNSS jamming or spoofing. That is a much bigger problem. Satellite navigation serves aircraft, emergency services and everyone living near the institution, and GNSS interference does not stop at the fence. Nothing public says whether CSC’s jammers operate in GNSS bands at all, and the order’s requirements for NAV CANADA notice and minimum-area, minimum-duration interference show how seriously ISED treats spillover from any jamming near an institution.

This is why the pilot’s results matter. If jamming is working, scale it. If drops continue because smugglers have shifted to autonomous routes, more jammers will not fix it. The answer then is layered defence: radar and electro-optical sensors alongside RF detection, fast ground response to drop points, physical barriers, and interdicting operators on the ground. Canada already has proof that ground response works: a Kingston task force linking Kingston Police, CSC and the Canada Border Services Agency cut local drone drops by roughly half in nine months, by police estimate. Yet The Walrus reports Stony Mountain Institution is set to lose detector dogs, mobile patrol vehicles and officer positions to federal budget cuts. The United Kingdom has gone the layered route: £35 million to fit up to 13,000 cell windows with steel grilles, on top of £10 million already spent on counter-drone measures such as external netting. Right now, the only people who know which of those is true are the ones running the pilot.

The Drone That Cannot Be Jammed

There is a harder problem coming, and Ukraine has already shown it to the world. Since 2024, both Russian and Ukrainian forces have fielded FPV drones controlled through a thin fibre-optic cable that unspools behind the aircraft. The control inputs and video travel as light through glass. Nothing is transmitted over the air.

That design choice defeats the two tools corrections is building around. A jammer has no radio link to disrupt, so the drone flies through it. RF detection has no signal to hear, so the drone is invisible to any sensor that listens for control or video transmissions. CBC reported in early 2025 that experts expect this to matter well beyond the war zone, precisely because jammers are one of the few tools security agencies have against rogue drones.

The barrier to entry is low. Underneath the spool, a fibre-optic FPV is built from hobby-grade parts: motors, frame, battery, camera, goggles. The spools themselves are commercially manufactured, with Chinese factories reported as supplying both sides of the war. Smaller battlefield models fly a couple of kilometres; larger ones reach about 20. A prison drop needs a fraction of that.

There is no public report of a fibre-optic drone being used against a Canadian institution. But smugglers adapt fast. U.S. corrections officials describe drops growing from about four pounds per flight to 25-pound loads on heavy-lift aircraft. A technique that defeats both jamming and RF detection will not stay on the battlefield.

It is not unstoppable. A fibre drone is heavier and slower than a radio one. Radar and electro-optical sensors can still see it. The cable can snag, and it leaves a physical trail back toward the launch point. But every one of those counters is a sensor or a physical response, not a jammer. A pilot program that measures only whether jamming works is evaluating the last war’s tool.

A Pilot Is Not a Policy

Here is the antagonist, and it is not a lack of law. It is governance by ministerial pilot — a narrow, temporary, spectrum-only answer to a threat that is already moving off the spectrum.

It is temporary. SOR/2025-63 expires on its fifth anniversary. The government itself describes it as a pilot designed to generate data for ISED’s future regulatory approach. Capability built around it has a sunset date.

It is narrow. The order covers CSC and Quebec’s correctional services. That is all. Provincial jails in Ontario, Saskatchewan, Manitoba and elsewhere remain under the full jammer prohibition. Saskatchewan announced in April 2026 that it would test its own drone detection technology for provincial facilities. Without an exemption order, detection is the legal ceiling for provincial corrections outside Quebec. ISED’s March 2026 Forward Regulatory Plan signals a separate proposal to exempt certain law enforcement agencies, but as of this writing no order covers provincial jails outside Quebec.

It is opaque. Nobody should expect CSC to publish where its jammers are or when they switch on. That would hand smugglers a map. But eighteen months in, there is also no published aggregate data: how many activations against drones, and whether drop rates changed at sites that use them. Results by region or fiscal year give nothing away, and they are exactly what the pilot exists to produce. When The Walrus examined the issue in August 2026, it found details of the jamming and detection pilots sparse, and CSC’s own statement pointed to detection systems and body scanners, not jamming.

It is spectrum-only. A Radiocommunication Act exemption covers only who may interfere with radio signals. Bill C-15 (Royal Assent March 26, 2026) now bars any counter-drone use without the Minister of Transport’s authorization, with details to follow in a Transport Canada drone security program.

The Decision That Has to Be Made

For CSC, the legal question is settled. The harder questions are not. Is the pilot working? Will provincial correctional services outside Quebec get the same authority, or keep filing detection reports? Will the evaluation measure the whole detect-and-defeat chain, including radar, physical barriers and ground response, or only whether jamming works? And when Transport Canada begins authorizing agencies to interdict drones under its new Aeronautics Act powers, will corrections be first in line?

Canada spent seven years building Part IX into a credible framework for lawful drones — as of August 13, 2026, Transport Canada counts 709 Level 1 Complex pilot certificates issued and 615 active RPOCs. It has given the drones flying contraband into its prisons an 18-month-old jamming pilot — while the drones that jamming cannot touch are already in serial production. One of those deserves the same seriousness as the other.