Publications

The most consequential choice is not always which surveillance tool a city buys. It is what happens when individually procured systems become one interoperable policing infrastructure.

A facial recognition result can look like a fixed algorithmic output. But some systems allow analysts to alter the candidate list before the result is preserved or disclosed.

Axon’s Draft One does not simply “watch” body-camera video and write a report. Understanding what the system actually receives, transforms, and generates matters to understanding the evidence it produces.

Scientific validation testing does not simply establish whether a method works. Done effectively, it defines where the method stops being valid. It tells us where to stop.

“AI” describes very different systems performing very different tasks. Before we can evaluate an AI system, we need to ask what data it uses, what task it performs, what it produces, and what claim is being made from that output.

DNA evidence still carries the authority of the “gold standard.” But increasingly sensitive testing has changed both the evidence we recover and the questions that evidence can actually answer.

Jail communication systems no longer simply transmit calls. They turn conversations, voices, visits, payments, and relationships into data that can be stored, searched, analyzed, and used by the state.

The newest surveillance tool is rarely the beginning of the story. Its power rests on an accumulated infrastructure of data, systems, and traces that was built long before the latest technology came into public view.