EFF Says Lawmakers Have “Precious Little Time” to Stop Armed Police Drones, and the Skydio Reversal Is Why

The Electronic Frontier Foundation published a warning on June 26 that armed police drones are no longer a hypothetical, and that the absence of any federal law governing when American law enforcement can deploy force from a drone has left the question to the companies selling the hardware. The digital rights group, which has argued against police arming robots and drones since 2021, pointed to two developments from this month as evidence the country may be sliding toward routine drone militarization. One of them was the Skydio CEO walking back a no-weapons pledge on a national podcast, a reversal DroneXL covered in detail eight days before EFF’s piece ran.

EFF’s core demand is narrow and specific. Cities should not buy weaponized drones or robots, and the multi-purpose machines departments already operate should be barred by policy from causing physical harm, whether that means deadly force or less-lethal options like pepper spray, kinetic strikes, rubber bullets, or Tasers. The group argues that relying on the goodwill of vendors is no longer enough, because the vendors have started saying the goodwill has limits.

I have spent the better part of a decade watching the police drone debate run almost entirely on surveillance grounds. The privacy fight over what a camera in the sky can see is familiar territory. Weaponization is a different argument, and it is arriving faster than the policy conversation around it.

Skydio’s CEO told a podcast that drawing weapons red lines is “dangerously misguided”

The first development EFF flagged is the one DroneXL reported on June 18. On a June 15 episode of The Verge‘s Decoder podcast, Skydio CEO Adam Bry distanced his company from a written commitment, still posted on Skydio’s own site, stating it would not put weapons on its drones. Bry told host Nilay Patel that the earlier messaging had led people to believe Skydio would prevent the military from mounting weapons on its aircraft, and that the company no longer holds that position.

EFF read the comments the same way DroneXL did, as a signal that Skydio will not restrict how its customers use its hardware. The group’s framing was blunt: whether police arm drones domestically currently rests more on the internal ethics of companies than on any law passed by an elected official. Bry’s own words on the podcast went further than a quiet policy retirement. He called the impulse to write such restrictions “dangerously misguided,” arguing that bright-line refusals produce adverse selection, because responsible buyers honor terms of service while adversaries and terrorists ignore them. It is a coherent argument. It also clears away every constraint on a defense contractor’s product roadmap, and it is being made by the CEO of that contractor.

This is not abstract. As DroneXL reported in June 2025, soldiers from the 278th Armored Cavalry Regiment dropped live M67 fragmentation grenades from a Skydio X10D at the Grafenwoehr training area in Germany, using a 3D-printed munitions dropper. Bry confirmed on Decoder that the Army has run those grenade-dropper experiments and that internal calls to shut them down did not win. The exact thing Skydio’s posted principle said it would not do has already happened on a Skydio airframe.

A company is putting suppression drones in Georgia and Florida schools this fall

EFF’s second development is a school-safety startup called Campus Guardian Angel, which will run pilot programs in Georgia and Florida high schools starting this fall. The Austin, Texas company, founded by Justin Marston and Bill King, builds what it calls an Active Shooter Suppression System: small drones stationed on charging pads throughout a building, flown remotely from the company’s headquarters, designed to reach an attacker in roughly 15 seconds and deploy less-lethal effects to buy time before police arrive.

Those effects are not gentle. The drones can blast sirens, fire strobe lights to disorient, spray pepper gel, and ram an attacker in a kinetic strike. Georgia allocated $550,000 for five high schools, and Florida appropriated $557,000 for the Broward, Leon, and Volusia districts, with Deltona High School first online. The Georgia pilot runs through the end of the 2026 to 2027 school year. Marston has said openly that the idea came from watching small drones used against soldiers in Ukraine, with the difference being that his drones are meant to incapacitate rather than kill.

DroneXL has covered this program closely, including Georgia’s $550,000 bet and the earlier Texas demonstrations, and the assessment has been consistent: the human-piloted design deserves real credit, and the certainty being sold around it does not. The hard numbers come from demos, and in demos everything works. EFF raised the same operational doubts from the civil-liberties side, noting that the public demonstrations mostly show drones crashing into stationary mannequins in pristine conditions, and asking what happens when a shooter fires at a moving drone with people fleeing behind it. The group also tied the program back to Axon, the Taser and body-camera maker that paused its own Taser-armed drone plan in 2022 after its ethics board resigned over the proposal.

Federal law governs shooting at drones, not drones shooting at people

The regulatory gap EFF is pointing at is real and lopsided. Federal law is clear and well-tested on one direction of force: firing at a drone is a felony. As DroneXL reported this month, a man who fired a BB gun at a Lee County Sheriff’s Office drone is facing charges under 18 U.S.C. § 32, which treats a drone as an aircraft and carries up to 20 years in prison. The statute dates to 1956 and has been applied to drone cases for years.

There is no equivalent body of law running the other way. No federal statute defines when a law enforcement drone may deploy force against a person, what counts as proportionate, or who is liable when a less-lethal effect injures a bystander. San Francisco became the first city to ban police from using deadly force via robot back in 2022, but EFF argues a single-city policy covering only robots, not drones, is not enough. A workable rule, the group says, has to cover both categories and explicitly prohibit any body harm. The asymmetry is the whole problem: the person on the ground who shoots at a police drone faces two decades in federal prison, while the drone that might one day deploy force against that person operates in a legal vacuum.

The DJI ban removed the cheap competition and handed the field to American vendors

The context EFF mostly leaves out is how thoroughly the procurement field has been cleared. The FCC placed DJI on its Covered List in a December 22, 2025 ruling that blocks new equipment authorizations, and DJI has filed a petition for review in the Ninth Circuit (Case 26-1029), estimating the ban will cost it roughly $1.56 billion in lost U.S. revenue this year. On Decoder, Patel put the dynamic to Bry plainly: the government handed Skydio a gift by removing the affordable competition that served first responders, and Skydio can now sell expensive programs to agencies with no other options. Bry did not dispute the framing.

That matters for EFF’s argument in a way the EFF piece does not develop. The same policy push that promises domestic security and supply-chain integrity is also concentrating the police drone market in the hands of a small number of American defense-adjacent vendors, at least one of which has just announced it will no longer draw a line at weapons. Bry conceded on the podcast that, at equivalent price points, China still builds better drones, and that even his own aircraft carry untraceable Chinese content past the first supply tier. The market is consolidating around companies whose ethical commitments are, by their own CEO’s account, negotiable.

DroneXL’s Take

I disagree with EFF on plenty. The group treats police drones as close to presumptively suspect, and I do not. I have spent years documenting drones pulling people out of rivers, finding lost kids in the dark, and clearing a building so an officer does not have to walk into a barricaded room blind. A Drone as First Responder program is one of the genuinely good things this technology does. That conviction is not in tension with the warning EFF is issuing. It is the reason to take the warning seriously.

Here is the line worth holding. Surveillance and weaponization are separate fights, and the second one is winnable right now in a way the first never fully was. There is no constituency demanding that American police drones carry pepper spray and kinetic strikers. No department is begging for it. The only pressure pushing in that direction is a vendor who decided that drawing red lines is bad for business, dressed up as a philosophy about adverse selection. A grenade dropper already flew on a Skydio airframe in Germany. Campus Guardian Angel is about to put ramming-and-pepper-spray drones inside high schools where teenagers go to class. The distance between “less-lethal effects in a school” and “less-lethal effects at a protest” is a procurement decision, not a technical barrier.

And the asymmetry should make every operator uneasy. If you fire a BB gun at a sheriff’s drone, the federal government can put you away for twenty years, and it should, because that drone is an aircraft and the people relying on it are doing real work. But there is no statute anywhere that tells a police drone what it may not do to the person on the ground. We built an entire legal edifice protecting the machine and left the human it might one day target completely uncovered. EFF wants lawmakers to close that gap before armed police drones become normal. On that specific point, the civil-liberties group and this pro-drone publication are in complete agreement, and the drone industry should want the line drawn too. A single department arming a drone and hurting a bystander would do more damage to public trust in this technology than every privacy complaint of the last decade combined. Don’t be the company, or the city, that finds out.

Watch the Florida and Georgia pilots through the 2026 to 2027 school year, and watch whether any state legislature moves a bill that actually defines the limits on drone-deployed force. The Campus Guardian Angel contracts have an end date written into them. Whether they get renewed, expanded, or quietly studied for what the less-lethal effects actually do to a moving target is the next real test, and it is one DroneXL will be covering.

Source: Electronic Frontier Foundation.

DroneXL uses automated tools to support research and source retrieval. All reporting and editorial perspectives are by Haye Kesteloo.