The Question

Smart glasses with embedded camera and microphone worn in a crowded public space

In October 2024, two Harvard students, AnhPhu Nguyen and Caine Ardayfio, demonstrated something that alarmed the internet. Using Meta's Ray-Ban smart glasses — a mainstream consumer product starting at $299 — and software they called I-XRAY, they looked at strangers and, within about a minute, pulled up names, home addresses, and phone numbers. The glasses streamed video; a face-search service matched the faces; data brokers supplied the rest.

The students said they would not release the tool; they built it to show the risk. But the hardware is selling fast. EssilorLuxottica, which makes the glasses with Meta, said about 2 million had been sold between late 2023 and early 2025 — and roughly 7 million smart glasses were sold in 2025 alone. The glasses have a small light that shows when they are recording, but critics say it is easy to miss. In a world where everyday eyewear can capture what a person sees and hears, what does "private conversation" mean any more?

What the Evidence Shows

The erosion of private conversation is not one technology but a convergence. Microphones are tiny and cheap. Wireless networks can stream audio from almost any device. AI transcription has become accurate and free: OpenAI's Whisper model, released as open source in 2022 and trained on 680,000 hours of audio, transcribes dozens of languages and can run on a laptop. Cloud storage costs only a couple of cents per gigabyte a month, so a year of continuous compressed audio costs a few dollars to keep. And consumer wearables have normalised the idea of always-on recording.

Smart speakers made ambient listening normal. They listen locally for a wake word, but they sometimes wake by mistake. In 2019, Bloomberg reported that Amazon employed thousands of people around the world to listen to and transcribe Alexa recordings to improve the service. A 2020 study by researchers at Northeastern University and Imperial College London found that smart speakers playing TV dialogue activated by mistake between 1.5 and 19 times a day; about half of these misactivations lasted less than six seconds, but some recorded as much as 43 seconds of audio.

"Modern cell phones are not just another technological convenience. With all they contain and all they may reveal, they hold for many Americans 'the privacies of life.'"

— Chief Justice John Roberts, Riley v. California, US Supreme Court, 2014

The more disruptive device class is wearables. Meta's camera glasses record photos, video, and audio. A new category of AI "memory" pendants and wristbands records conversations all day and summarises them: Amazon bought one maker, Bee, in July 2025, whose $50 device records all audio within range unless muted, and Meta bought another, Limitless, in December 2025. Meanwhile AI note-takers now join video meetings by default in many workplaces. Not every product succeeds — Humane's AI Pin was discontinued in 2025 — but the direction is clear.

The secret is not dying — it has already died. We just haven't updated our social norms to acknowledge the body.

Why This Is Happening

Economics made it inevitable. The cost of capturing, transmitting, and storing audio has collapsed. Surveillance of communications once required nation-state resources, as Edward Snowden's 2013 revelations about NSA programmes showed. Today, a comparable ability to record and search what people say nearby is available to any individual with a few-hundred-dollar device and a cloud subscription.

AI transcription removed the last bottleneck. Recording audio was always possible; surveillance stayed limited because humans had to listen to extract meaning, which was expensive. Whisper and similar models changed that, transcribing speech faster than real time at no per-use cost. An hour-long conversation can now be transcribed, summarised, and searched in minutes. Audio is becoming as searchable as email.

Social normalization is accelerating adoption. People already wear earbuds in almost every social setting. Apple's AirPods Pro have a "Conversation Boost" mode that amplifies nearby speech, and in 2025 Apple added live translation of in-person conversations. When the devices people wear every day routinely process everything spoken near them, the psychological barrier to ambient recording falls.


What Could Happen

Normalization without legislation Most likely · 55%

Recording technology proliferates faster than law. By 2032, most professional and social interactions in developed countries are recorded by at least one party's wearable device as a routine matter, much as smartphone cameras normalized photography of previously private moments. Society adapts through informal norms — "can I record this?" becomes a standard social courtesy — but legal frameworks lag by a decade, creating a period of profound ambiguity for whistleblowers, therapy patients, attorneys, and anyone engaged in sensitive communication.

Regulatory crackdown creates two-tier world Possible · 30%

The EU and some US states impose strict ambient recording rules — mandatory indicator lights, consent requirements, or limits on facial recognition. About a dozen US states already require the consent of everyone in a conversation before it is recorded. This creates a divergence: tighter protection in some places, few limits in others.

Counter-technology arms race Less likely · 15%

A market for anti-surveillance technology emerges: microphone jammers, acoustic "privacy rooms", and devices that detect recording wearables. A prototype "bracelet of silence" developed by University of Chicago researchers in 2020 uses ultrasound to jam nearby microphones. The result is an uneasy stalemate rather than a durable resolution.

Our Assessment
We assign 80% probability — likely that by 2035, recording and transcribing conversations will be cheap, ubiquitous, and socially commonplace. We have lowered this forecast slightly from 88%. Our earlier version misdated the Harvard demonstration, misstated product features and sales, and cited an Electronic Frontier Foundation quote and statistics we could not verify. The core trend is real: camera glasses are selling millions, Amazon and Meta have bought always-on recording start-ups, and AI note-takers are routine at work. The key uncertainty is social and legal — consent laws and indicator rules may make recording more visible, but they are unlikely to stop it.

What Can We Do

Person speaking privately in a shielded, technology-free room

The erosion of conversational privacy is not fully reversible, but its consequences are not predetermined. Individuals, institutions, and policymakers each have meaningful responses available.

Demand consent indicators on all recording devices. Visible recording lights and clear consent rules make ambient capture transparent. Supporting laws that require them — and choosing devices with obvious indicators — creates market pressure for transparency.

Designate genuinely private spaces. Hospitals, therapy practices, law firms, and houses of worship can set technology-free conversation zones, as some courts and exam halls already do.

Understand what you already agreed to. Smart speakers, phone assistants, and AI note-takers store recordings under terms most users have never read. Review your Alexa voice history and Google Assistant activity, and check the recording settings of any meeting software you use. The recordings may already exist — the question is whether you know about them.

Push for stronger whistleblower and therapy privilege laws. The most serious harms from ambient recording will fall on the most vulnerable: abuse survivors confiding in counselors, employees reporting corporate wrongdoing, journalists protecting sources, attorneys advising clients. These conversations deserve categorical legal protection regardless of the technical means used to capture them — and current law does not adequately provide it.