Criminal at install time

Sam Tunick landed at Hartsfield-Jackson on January 24 of last year, back from the Dominican Republic. Agents pulled him into secondary screening and asked him to unlock his phone. They asked repeatedly, and warned they would seize it if he refused. When he finally entered a passcode, according to his defense motion, the screen went blank, flashed several times, and the phone appeared to restart. The Justice Department has charged him under a federal statute that makes it a crime to destroy property to prevent its seizure.

The charge asks a court to treat a configured default as an act. That is the part worth watching, and it does not depend on whether Tunick turns out to be sympathetic.

GrapheneOS runs on Pixel phones and supports a duress passcode: a second code that, when entered, wipes the device. You set it up once, in a settings screen, on a quiet afternoon. It then sits there indefinitely doing nothing. The wipe is not an act you perform in a screening room — it is an act you arranged months earlier and, under compulsion, did not stop. Prosecutors are collapsing that distance. Researchers quoted in the reporting said they had not seen the statute aimed at an operating system before.

That is Christophe Boutry, a cybersecurity and surveillance researcher; Bill Buddington of the Electronic Frontier Foundation said the same. The context sharpens it. Court testimony indicates federal agents had already circulated Tunick's name and photo internally, describing him as under investigation for "suspected terrorism activities" on the basis of alleged association with the movement opposing Cop City, the $109 million Atlanta police training facility that opened last spring. His filing says he asked for a lawyer four times and was refused each time, and that no warrant was produced.

The government's best argument is not weak. He chose which code to type. A duress passcode deliberately entered while agents demand access is a decision, and pre-arranging the mechanism does not launder the intent behind pulling the trigger on it. Automating the execution of a choice is still the choice. Any engineer who has hidden a destructive operation behind a config flag understands exactly why that reasoning holds.

But that argument requires proving what he typed and that he knew what it would do, and it is not what makes this case new. What makes it new is that the theory keeps reaching past the man and landing on the tool. Boutry points to Catalonia, where police have profiled people carrying Pixel phones on the assumption that a Pixel means GrapheneOS and GrapheneOS means a dealer. Nobody there had to type anything.

The line between a security control and destruction of evidence is not in the code. It is in who the government has already decided you are.

This is not an exotic feature under an exotic threat model. Corporate MDM wipes a lost laptop on command, and every IT department in the country treats that as hygiene. Apple ships an Erase Data option that clears an iPhone after ten failed passcode attempts. Full-disk encryption makes a drive unreadable by design, which is the entire point of buying it. All of these destroy the government's access to data in a manner arranged well in advance. The only thing separating a compliance checkbox from a felony charge is the file the agents were already carrying when you walked up to the counter.

A judge is not expected to rule on the suppression motion until late October. Until then the operative claim stands: that a man committed destruction of evidence by entering a number into a phone that had been told, long before, what to do with it. If that holds, the crime was committed at install time.