Field Notes

source

Mozilla anti-surveillance fashion tests

StatusStableUpdatedStands on2 sourcesSupportsnotes

Mozilla Foundation’s Nothing Personal series published a review and a feature on anti-surveillance fashion in November 2025. They are the most useful independent product tests of commercial Adversarial clothing located so far.

What was tested

The review covered garments from AntiAI Clothing, Yelo Pomelo, Cap able, Oogaly, and RiotAssembly, plus Reflectacles IR eyewear, at prices from USD 8 to 770. Test rigs were an Imou consumer security camera with person-detection alerts, iPhone Face ID, Windows Hello, and a pose-estimation model built with Google’s Teachable Machine.

Findings

  • All tested AntiAI and Yelo Pomelo garments defeated the Imou camera’s “human detected” alert when the reviewer walked past.
  • The same garments failed against the pose-estimation model, confirming that a pipeline change recovers detection.
  • Reflectacles glasses did not affect person detection but defeated Face ID and Windows Hello enrollment and login.
  • Cap able validates its knitwear against YOLO, an object detector, which says nothing about production recognition pipelines.
  • AntiAI’s founder reported over 7,000 orders in three years, growing European interest, and acknowledged the algorithms are getting harder to trick.
  • The feature frames the garments primarily as awareness and protest objects, citing researcher Torin Monahan’s phrase “aestheticization of resistance” and Adam Harvey’s warning that technical camouflage is temporary.

Use and limits

These are single-rig tests by a foundation with an advocacy mission, not laboratory benchmarks. They show what specific products did against specific consumer systems on specific days; they do not validate any garment against deployed police or commercial FR. The price and seller data informed Anti-surveillance apparel market.

Built on 2 sources (2 external).

Working out connections…