In an egregious example of the growing risks posed by camera-equipped wearables, an August 2026 Boston police report alleges that a registered sex offender used smart sunglasses to film children at a public spray pool (see go.nature.com/471zlwm). The incident illustrates how such devices can subject bystanders to intrusive surveillance without their knowledge or consent. The scale of the problem will only be compounded as the devices become integrated with artificial intelligence and the market expands.

Don’t sleepwalk from computer-vision research into surveillance

Technology firms are rapidly turning smart glasses from a niche gadget into a mass-market product, selling eyewear fitted with sensors as a tool for convenience and a platform for AI integration. They are the fastest-growing wearable category. More than eight million smart glasses were shipped in 2025, and by June 2026, sales of AI-equipped smart glasses had reportedly risen by more than 700% compared with the previous year. As someone who has spent a decade studying the potential harms associated with AI-powered biometric technologies, I urge policymakers to prepare for the major privacy-related crisis that is brewing.

Even at this nascent stage of adoption, media reports of bystanders expressing worry about ‘pervert glasses’ show why camera-equipped wearables pose a distinct problem: they can be more discreet than phone cameras and more portable than security cameras. These devices also mark a decisive shift from AI being confined to screens to it being embedded in consumer products that can continuously capture the physical world around us. AI-enabled wearables, including glasses, headphones and pendants, could thus expose millions of people to surveillance at an unprecedented scale.

Policymakers must therefore act now to update regulatory frameworks, before ambient AI becomes ubiquitous. They should include the creation of ‘consent-conscious spaces’ — such as schools, hospitals, courts and swimming pools — in which covert recording is prohibited by law and meaningful penalties are introduced to deter violations.

Halt the use of facial-recognition technology until it is regulated

A 2024 project by students at Harvard University in Cambridge, Massachusetts, which anticipated the current wave of wearable devices, showed the inherent risks in pairing AI with smart glasses (see go.nature.com/3tapz0y). In the demonstration, smart glasses live-streamed video to social media, on which an AI-powered facial-recognition system analysed the feed. The students said their aim was to show how existing tools can be combined with discreet recording devices to uncover sensitive personal information. A faceprint — a digital map of a face — created from a person’s image can be matched against databases of previously collected faces. Once it is linked to a name in a database, other personal details, including home addresses, can also be exposed.