Do Courts Need to Start Buying AI Jammers?
Late last year, I wrote about the risk of AI wearables walking into court. At the time, the issue still had the feel of something just over the horizon. It was easy to imagine many judges reading about smart glasses, invisible augmentation, quiet recording, and real-time assistance and treating it as one more developing technology issue that would need attention someday, but not necessarily today. That is often how these transitions begin. The technology first appears in stories that sound novel or niche, and only later does it become clear that the shift was already underway.
Then came the report that Mark Zuckerberg walked into court and members of his entourage were wearing Meta glasses. The reason that story landed so hard is not that it was shocking. It is that it felt like immediate confirmation of a concern that had already been raised. In that courtroom, the judge had cues. It was a Meta case and Zuckerberg was physically walking into the building. The connection between the technology and the proceeding was close enough that suspicion came naturally. In other words, the issue announced itself.
Most judges will not get that kind of warning.
The next generation of AI wearables will come in ordinary form, and that is what makes the problem harder. By the time the technology blends into the background, the court may not recognize the issue until after the damage is done.
That thought stayed with me when I later read about a recent Harvard graduate developing an AI jammer, a device reportedly designed to interfere with nearby microphones and wearable recording systems. My initial reaction was simple: do courts need to start buying jammers? The more I sat with that question, the more revealing it became. If people are already trying to build countermeasures for ambient recording and wearable surveillance in everyday life, courts should at least ask what that means inside a courtroom, where witness safety, sequestration, and the integrity of testimony matter in a very different way.
One part of the concern is visual, and that is where smart glasses become the clearest example. In a murder trial, or in any case involving intimidation, retaliation, or fear, the risk is not limited to someone brazenly pulling out a phone and recording a witness. The more realistic concern is a person sitting quietly in the gallery wearing ordinary-looking glasses capable of video recording that witness without drawing immediate attention. That kind of recording is not just a privacy issue. In the wrong case, it can become a witness-safety issue with real consequences. The same risk extends to jurors. A device capable of recording faces in a courtroom today may be capable of identifying them tomorrow. In the Meta case, Judge Kuhl went beyond ordering the glasses removed and any footage deleted if it existed. She specifically banned any use of facial recognition technology to identify jurors. That order was possible because the threat was visible. In most courtrooms, it will not be.
But the harder problem, at least from a courtroom management perspective, may be audio. Sequestration and witness coaching are discreet-audio-device problems. That is where the technology becomes more difficult for courts to identify and much harder to regulate through traditional assumptions about what a prohibited device looks like. No one is likely to tolerate an obvious earbud in open court. The real challenge is the device that appears ordinary, unobtrusive, or even medically necessary. It is the hearing aid that is an actual hearing aid but may also do more than hearing aids once did. It is the lapel pin, the pendant, the wedding ring, and eventually the contact lens. The point is not that every one of these devices already poses the worst-case scenario. The point is that this is the direction of travel, and the courtroom’s old visual cues are becoming less reliable with each iteration.
That matters because courtroom procedure still assumes a world in which the meaningful technology in the room is visible enough to be recognized and controlled. A sequestration order still means what it has always meant in doctrine, but its practical force changes if testimony can be captured, transcribed, and relayed outside the courtroom through a device no one notices. A witness can still take the oath and sit on the stand, but the integrity of that testimony changes if real-time coaching or answers can be received through a device that appears entirely unremarkable from the bench.
This is why the jammer story struck me as useful, even if the answer for courts is not going to be some simple piece of counter-surveillance hardware. The value of that story is not in the gadget itself. It is in what the gadget represents. It reflects a growing public assumption that ambient recording and invisible capture are becoming ordinary features of modern life. Once people start imagining tools to defend against that reality, it is a sign that the underlying concern has moved beyond theory. And if it has moved beyond theory for ordinary life, courts should not comfort themselves with the idea that they still have time to wait.
The judge in the Meta case had cues that made the issue easier to spot. Most judges will not be so lucky. In most courtrooms, the problem will not announce itself. It will sit quietly in the room, looking like a hearing aid or a ring, while carrying capabilities that courtroom procedure was never designed to confront in that form.
The more unsettling question is what it means that the jammer question now sounds plausible at all. That may be the more important signal, because it suggests that courts need to grapple much more seriously with the fact that the relevant technology is already becoming harder to see, harder to categorize, and harder to separate from ordinary human presence.
