by Sarah Elaine Eaton – March 18, 2026
A London judge recently concluded that a witness was receiving coached answers through a pair of smart glasses connected to his mobile phone during cross-examination (Jacobs, 2026). The case involved a routine insolvency dispute, but the technology at the centre of the judge’s findings was anything but routine. The witness, who gave evidence through a Lithuanian interpreter, was found to have been receiving audio from an unidentified caller routed through smart glasses paired to his handset. Once the glasses were removed, his phone began broadcasting a voice from his jacket pocket. The judge rejected the witness’s testimony in full, describing it as unreliable and untruthful.
The incident is instructive for those of us working at the intersection of technology, integrity, and institutional policy. It demonstrates that smart glasses do not need advanced AI capabilities to compromise a formal proceeding. Simple Bluetooth audio connectivity was sufficient.
In our recent paper (Eaton et al., 2026), we examined the implications of AI-enabled smart glasses for teaching, learning, assessment, and academic integrity. One of our central arguments applies here: the reflexive instinct to treat wearable technology as a cheating device, while understandable, risks missing the structural challenge these technologies present to the systems designed to ensure honest participation.
Courts, like universities, depend on observable behaviours and verifiable evidence to assess credibility and ensure procedural fairness. As we noted, AI glasses can embed cognitive or communicative assistance into a user’s perceptual field in ways that leave no external trace (Eaton et al., 2026). The London case illustrates what happens when that assistance leaves a trace, but only because something went wrong: the interpreter heard voices, and the phone began playing audio at the wrong moment.
The question this case raises is not whether courts should ban smart glasses. A blanket prohibition would create its own problems, particularly for individuals who depend on wearable technology for vision correction or accessibility. We argued that institutional responses should focus on redesigning processes rather than policing devices (Eaton et al., 2026). For courts, this means developing protocols for the use of wearable technology during testimony, much as we recommended that educational institutions establish centralized accommodation protocols for AI-enabled devices.
The London ruling also reinforces our observation that enforcement models built around detection are fragile. The coaching was discovered through a combination of the interpreter’s alertness, call log records, and the witness’s inability to explain the contact saved as “abra kadabra” on his phone. These are investigative tools, not systemic safeguards. As smart glasses become more common and more discreet, relying on detection alone will prove insufficient in both courtrooms and classrooms.
What this case calls for is not alarm but preparation. Institutions responsible for the integrity of formal proceedings, whether legal or academic, need forward-looking frameworks that address the capabilities of wearable technology before the next incident occurs. The technology is not going away. Our systems must adapt.
References
Eaton, S. E., Kumar, R., Dahal, B., Tang, G., Ramazanov, F., & Moya Figueroa, B. A. (2026). AI smart glasses and the future of academic integrity in a postplagiarism era. Canadian Perspectives on Academic Integrity, 9(1), 1–5. http://doi.org/10.55016/ojs/cpai.v9i1/82885
Jacobs, S. (2026, March 17). A London judge says a witness was being coached in real time through smart glasses. TechSpot. https://www.techspot.com/news/111710-london-judge-witness-coached-real-time-through-smart.html
____________
Cross posted from:
From Courtrooms to Classrooms: Smart Glasses and Integrity in a Postplagiarism Era – https://postplagiarism.com/2026/03/18/from-courtrooms-to-classrooms-smart-glasses-and-integrity-in-a-postplagiarism-era/
Learning, Teaching and Leadership 