The prospect of American skies swarming with drones raises more than just safety concerns. It alarms privacy advocates as well. Infrared and radio-band sensors used by the military can peer through clouds and foliage and can even—more than one source tells me—detect people inside buildings. Commercially available sensors too are extraordinarily sensitive. In Colorado, Chris Miser detaches the infrared camera from the Falcon, points it at me, and asks me to place my hand on my chest for just a moment. Several seconds later the live image from the camera still registers the heat of my handprint on my T-shirt.

During the last few years of the U.S. occupation of Iraq, unmanned aircraft monitored Baghdad 24/7, turning the entire city into the equivalent of a convenience store crammed with security cameras. After a roadside bombing U.S. officials could run videos in reverse to track bombers back to their hideouts. This practice is called persistent surveillance. The American Civil Liberties Union (ACLU) worries that as drones become cheaper and more reliable, law enforcement agencies may be tempted to carry out persistent surveillance of U.S. citizens. The Fourth Amendment to the Constitution protects Americans from “unreasonable searches and seizures,” but it’s not clear how courts will apply that to drones.

What Jay Stanley of the ACLU calls his “nightmare scenario” begins with drones supporting “mostly unobjectionable” police raids and chases. Soon, however, networks of linked drones and computers “gain the ability to automatically track multiple vehicles and bodies as they move around a city,” much as the cell phone network hands calls from one tower to the next. The nightmare climaxes with authorities combining drone video and cell phone tracking to build up databases of people’s routine comings and goings—databases they can then mine for suspicious behavior. Stanley’s nightmare doesn’t even include the possibility that police drones might be armed.