Is It Legal to Record Strangers With Smart Glasses? The Law Asks What They Expected

Every Test the Law Applies Depends on Them Knowing
The question gets asked constantly and the answers are remarkably consistent. Silent video in a public place is generally fine, because nobody there has a reasonable expectation of privacy. Audio is a different statute, and in some states recording a private conversation without everyone's permission is a crime. That is the shape of every guide on the first page of results, and as a summary of the law it is broadly right.
It is also answering a slightly different question from the one being asked, and the gap only shows up when you read the statutes rather than the summaries of them.
What the federal statute actually protects
The one-party consent rule everyone cites is real and it is narrow. Section 2511(2)(d) of the federal wiretap act says it is not unlawful for a person not acting under colour of law "to intercept a wire, oral, or electronic communication where such person is a party to the communication or where one of the parties to the communication has given prior consent to such interception."
Read that clause for what it requires rather than what it permits. It covers a person who is in the conversation, or who has one participant's permission. A stranger walking past two people talking is neither.
Then there is the definition of the thing being protected, which is where the interesting part is. Section 2510(2) defines an "oral communication" as "any oral communication uttered by a person exhibiting an expectation that such communication is not subject to interception under circumstances justifying such expectation."
That is not a fixed category. It is a test with two halves, and the second half — under circumstances justifying such expectation — is measured against the world as it is. What a person may justifiably expect in a café depends on what recording devices are ordinarily present in cafés. The statute does not say a conversation in public is protected. It says a conversation is protected if the speaker expected not to be intercepted and the circumstances made that expectation reasonable.
Which means the protection is not a wall. It is a reading of ordinary experience, and it moves when ordinary experience moves.
The count that is not a count
The other number in every guide is how many states require everyone's permission. It is worth noticing that the guides do not agree on it.
A health-law practice advising employers and clinics puts it at thirteen. A recording-law reference lists twelve, and the state on one list and not the other is Michigan.
The disagreement is not sloppiness. Michigan's eavesdropping statute reads as all-party on its face, and the exception that makes people call it a one-party state hangs on four words in a definition — "the private discourse of others". In Sullivan v. Gray in 1982, the state's Court of Appeals held that this reference to the conversations "of others" meant the statute could not apply to a person's own conversations. A participant is not eavesdropping, because a participant is not listening in on somebody else.
Follow that reasoning to the case people are actually asking about and it runs the other way. The same source is explicit that "it would be a mistake to believe a person could legally record a conversation in Michigan just because they were given permission by one of the participants." The exception belongs to participants. Somebody filming strangers on a street is not one, and for that person the shorthand disappears.
So the map is stricter for the bystander case than the headline number suggests — and the headline number is what a reader carries away.
It is worth being clear about what this does and does not show. Two careful sources counting the same thing arrived at different totals because one of the entries is genuinely contested, not because either was careless. The lesson is not that the guides cannot count. It is that a single number cannot carry a rule made of a statute, a 1982 appellate reading of four words inside it, and a distinction between participants and everybody else — and the number is the only part that survives being repeated.
Video and audio are not one decision
The guides are right that these are separate questions, and there is a reason it matters more with this device than with a phone.
The recording-law reference states the split plainly: silent video of a public scene "is governed by privacy and voyeurism rules and is usually allowed where there is no reasonable expectation of privacy," while the audio of a private conversation "is governed by your state's wiretap or eavesdropping statute."
The health-law advisory adds the part that closes the gap. These devices, it says, capture audio alongside video, as most do by default — and the moment they do, "the entire conversation is pulled back into the wiretap domain." Wiretap law is older than privacy-in-public doctrine and considerably less forgiving.
That is the trap in the plain-English version of the rule. "Video in public is fine" is true of video. A wearer who has decided, reasonably, that pointing a camera down a street is allowed has not made a decision about the microphone, because on this device the microphone was never a separate decision to make.
Every one of these tests needs the same missing thing
Line the tests up and they have a common dependency.
All-party consent requires the other parties to be asked, which requires them to know there is something to be asked about. The federal definition of a protected conversation requires an expectation that circumstances justify, which is a claim about what a person could reasonably have noticed. Even the public-place doctrine rests on what an ordinary person would understand about being observed.
Every one of them is keyed to the awareness of the person being recorded. And this desk spent a week establishing that the awareness is not available: every published method for spotting a recording carries its own admission that it fails, and the one detection app with radio behind it reports that a device is present, not that it is running.
The institutions with something to lose reached the same conclusion in their own way. DEF CON's stated reason for banning the device rather than issuing guidance was that there is no consistent way to tell whether a pair is recording, and the ban makes no exception even for prescription lenses. An employment-law practice writing for retailers says a venue rule of this kind rests on the honour system, for mechanical reasons: the device cannot be remotely disabled, what it is running is invisible to everybody but the wearer, and unlike a phone held up in plain sight it is far less conspicuous.
None of that makes the statutes unclear. The text is clear. What is missing is the fact the text depends on.
What that leaves
It leaves a legal question with a clean answer and an evidentiary one without.
The wiretap statutes were drafted around a bug in a room or a tap on a line — a thing installed deliberately, discovered afterwards, and provable. They work well there. Applied to a device that looks like eyewear and travels wherever its owner goes, the elements stay the same and the proof does not: whether a conversation was intercepted, by whom, and whether the speaker's expectation was justified at the moment they spoke.
And the second half of that federal definition is the part worth watching, because it is the only one that can quietly change without anybody amending anything. If undetectable cameras become an ordinary feature of public places, the circumstances that once justified expecting not to be recorded stop doing so on their own. The statute would not have moved. The world underneath it would have.
That is the most direct explanation for a pattern this desk has been tracking all week. Four institutions banned the device outright inside seven months, and the manufacturer has spent its effort on making the recording light impossible to defeat rather than on the question the bans named. A venue ban is a blunt instrument and everybody involved knows it. It is also the only instrument on the table that does not require the person being recorded to know they are being recorded — which is precisely what every one of the legal tests, and every one of the detection methods, turns out to need.
Sources
- 18 U.S.C. § 2511 — the federal wiretap act, including the one-party consent clause at (2)(d)
- 18 U.S.C. § 2510 — the definition of "oral communication" and the expectation test
- Dickinson Wright: wiretap exposure, audio capture by default, and workplace duties
- Recording Law: smart glasses recording laws, and the video-versus-audio split
- National Law Review: Michigan's participant exception and what it does not cover
- Digital Media Law Project: Michigan recording law
- Ogletree Deakins: why a venue rule rests on the honor system
- Cybernews: DEF CON's stated reason for banning rather than advising
- The Register: the ban, prescription lenses included
- Purdue Global Law School: smart glasses and privacy risk