AI Wearables Put Bystanders Under Surveillance

Chris Thomas wears Meta camera-equipped glasses, a Fieldy note-taking pendant and an Amazon Bee wristband to record parts of his daily life. At a local bar, a bystander told him, “I don't feel comfortable.” Thomas said he takes the glasses off completely because people “don't really believe it's off.” The devices promise convenience to their owners, but people nearby may get no say at all.

The wearer chooses; bystanders don't

Thomas, a 43-year-old Pittsburgh-based business owner and livestreamer, uses the devices to stay organized, pay bills on time and set reminders. “I'm forgetful AF,” he said. “My life is so crazy that not having to remember to remember is fantastic.” That bargain works for the wearer. For someone whose conversation or image enters a device's sensors without their knowledge, it can look very different.

Amazon's Bee Frontier wearable and Bee Eclipse, a card-shaped device that attaches to a smartphone, arrived Thursday. Both are designed to record, transcribe and analyze a user's day. Apple's latest Apple Watch Series 12, unveiled in September, will include Siri Recap, which “ambiently” takes notes throughout the day and produces conversation summaries. Live Rewind can recap 15 seconds of a conversation through a text transcription. Meta's smart glasses can take photos, capture video and record voice notes or reminders. Startups Plaud and Pocket have developed card-sized note-taking devices, and Friend offers an AI necklace.

The central question isn't only what the owner agreed to. What about people who didn't buy the device, accept the terms or choose to have a private conversation summarized? “A non-wearer did not buy the watch, agree to Apple's terms, or choose to have their conversation summarized in the app,” said Colin Zick, a privacy and data security attorney at Foley Hoag.

The law meets the apparatus

Earlier smart speakers usually relied on a “wake word” and sometimes stored recordings or sent audio to the cloud. Some newer wearables can continuously listen, interpret speech and extract meaning without retaining the underlying audio. Whether that processing itself can count as unlawful interception remains largely untested.

Many state wiretap and eavesdropping laws focus on whether someone intercepted a private communication or acquired its contents, not simply whether they made an audio file. Massachusetts defines an intercepting device to include equipment capable of “transmitting, receiving, amplifying, or recording” a communication. The communication's “contents” can include its substance and meaning. Zick said a device that continuously receives speech and produces a written summary could potentially fit that language even if it never saves an audio file.

Daniel Pi, an associate professor at the University of New Hampshire Franklin Pierce School of Law, said New Hampshire's statute is broad enough that he believes the Watch's audio-intelligence features would qualify as interception. The law covers recordings and the “aural or other acquisition” of a communication's contents. A consumer who enables such a feature during a private conversation without everyone's consent could, at least technically, violate state law.

Pi said a user in New Hampshire could potentially face civil or criminal liability, though he stressed the practical risk of prosecution is far lower. Establishing whether someone enabled the feature could prove difficult, and he said he would be surprised to see prosecutors pursue a bare statutory violation where no one was harmed. Earlier cases often involved devices mistakenly activating or capturing speech not intended as a command. These wearables are designed to acquire and interpret conversation, and users deliberately opt in; Pi said that difference could shift responsibility toward the wearer rather than the technology company.

Apple says Siri Recap doesn't store audio, send recordings to its servers or involve human review. Amazon says it doesn't store audio from Bee devices and that conversations are processed in a secure environment “where no one, including Amazon, can access your data.” Those assurances describe the companies' stated practices; they don't settle whether continuously interpreting a conversation meets the law's definition of interception.

A countermeasure for people in range

Aida Baradari, founder of startup Deveillance, created the hockey puck-shaped Spectre I for people concerned about recording without consent. It uses AI to scan for nearby devices and emits “adversarial signals” that scramble speech before microphones can process it. Baradari said she watched more AI wearables enter the market and felt users should have a greater say in how they interact with them. “We need to be able to use technology while not giving up our core values or rights,” she said.

Michael Stokes, a former CIA officer and senior vice president of emerging technologies at Veilant, said people can be swept into AI-powered surveillance without owning or knowingly interacting with the device collecting information. “I think the bystander problem is really the concern I have — that you can be surveilled without owning the technology,” he said. AI can rapidly interpret information gathered by cameras, microphones and sensors in devices such as smart glasses, watches and AI recorders, he said.

Stokes said the cost of making sense of collected information has fallen. Work that once could require thousands of hours to gather, sort and analyze information about a person is increasingly automated by AI. People who never opted into the technology can be exposed when a nearby wearable picks up a private conversation or a device they don't own collects information. “We no longer have these pockets of small privacy that we thought we did in the past,” Stokes said.