Is It Legal to Record a Conversation in Washington?

Statute · Penalty: mixed

RCW 9.73.030 (Washington Privacy Act, ch. 9.73 RCW — "Intercepting, recording, or divulging private communication—Consent required—Exceptions"); related: RCW 9.73.050 (inadmissibility), RCW 9.73.060 (civil action/damages), RCW 9.73.080 (criminal penalty), RCW 9.73.090 (law-enforcement recording exceptions)

In-person vs. phone & video calls

No meaningful split. RCW 9.73.030(1)(a) (telephone/telegraph/radio/other transmitted communications) and (1)(b) (in-person "private conversation" captured "by any device electronic or otherwise") both require consent of ALL participants — Washington does not default to one-party consent for either category. The only one-party-consent carve-outs are narrow, subject-matter-based exceptions in subsection (2) — emergency calls (fire/medical/crime/disaster), extortion/blackmail/bodily-harm threats, anonymous/repeated/inconvenient-hour calls, and hostage/barricaded-person communications — plus a bona fide news-media exception in (4), not a general phone-vs-in-person distinction. Subsection (3) lets a party satisfy "consent" for all others simply by announcing, in a reasonably effective and itself-recorded manner, that recording is about to occur — relevant to a visible/announced wearable device.

Recording in public

Washington's Privacy Act does not have a codified "public place" exception in the statutory text; instead, coverage turns entirely on whether the conversation was "private" in the first place. Washington courts assess privacy by the "intent or reasonable expectations of the participants as manifested by the facts and circumstances of each case," looking at duration and subject matter of the conversation, location, the relationship of the parties, and the potential presence of third parties. State v. Clark, 129 Wn.2d 211, 224-27, 916 P.2d 384 (1996) (drug-sale conversations conducted on public streets, in front of or within earshot of third parties/passersby, held not private, so recording them did not violate RCW 9.73.030). Conversations in public, especially routine transactional exchanges with strangers or within earshot of bystanders, are generally not "private conversations" and thus fall outside the statute's consent requirement — directly relevant to a wearable used to capture street disputes or interactions with strangers.

Recording at home

RCW 9.73.030 does not contain a distinct statutory rule for the home; the same "private conversation" test (State v. Clark) applies regardless of location, with location simply being one factor bearing on whether an expectation of privacy was reasonable. In practice, conversations occurring inside a private residence are much more likely to be found "private" (and thus require all-party consent to record the audio) than conversations conducted on a public street. There is no carve-out allowing a resident to make one-party-consent audio recordings of visitors/occupants in their own home; adding an audio track to a home security or doorbell camera can trigger the all-party-consent requirement (and potential gross-misdemeanor liability plus civil damages under RCW 9.73.060) if it captures a visitor's private conversation without consent, even on the recorder's own property. (This point about doorbell/security cameras specifically is drawn from consistent practitioner secondary commentary rather than a single controlling appellate case, so treat the doorbell-camera application as somewhat lower confidence than the general home/public-place framework.)

Recording the police

There is no statutory law-enforcement exception in Ch. 9.73 RCW that specifically authorizes a private citizen to record police (RCW 9.73.090 instead governs law enforcement's OWN body-camera/dash-camera and custody recordings, a different fact pattern). Instead, controlling Washington case law holds that an on-duty officer performing official public duties in the presence of a citizen, where the officer knows or should know the exchange may be observed, has no protectable "private conversation" interest — so recording that exchange is not a Privacy Act violation. State v. Flora, 68 Wn. App. 802, 845 P.2d 1355 (Div. I 1992), review denied, 121 Wn.2d 1028 (1993) (defendant's covert tape recording of police statements made during his own arrest held not a "private conversation"; court declined to let the Privacy Act become "a sword" for officers against citizens who record them). This rests on the same public-place/reasonable-expectation analysis as State v. Clark, above. Separately, on First Amendment grounds: Washington is within the U.S. Court of Appeals for the NINTH CIRCUIT. Fordyce v. City of Seattle, 55 F.3d 436 (9th Cir. 1995), recognized a First Amendment right to record matters of public interest in public places, including law enforcement officers performing official duties in public. Later Ninth Circuit authority, e.g., Askins v. U.S. Dep't of Homeland Security, 899 F.3d 1035 (9th Cir. 2018), reinforces a First Amendment interest in photographing/recording on publicly accessible property, though (per secondary commentary, e.g., Cardozo ICLR and RCFP surveys) the Ninth Circuit's doctrine on a "clearly established" right to record police is regarded as somewhat less definitively settled than in some sister circuits (e.g., 1st Cir. Glik v. Cunniffe). Net effect for the use case: recording an on-duty officer performing public duties in a public/quasi-public setting is very likely lawful under both the state Privacy Act (Flora/Clark line) and the First Amendment (Fordyce), but a private, non-public encounter with an officer could still be treated as a "private conversation" requiring consent absent one of the statutory exceptions or the announcement mechanism in RCW 9.73.030(3).

Video-only capture (no audio)

By its text, RCW 9.73.030 restricts recording of "communication[s]" and "conversation[s]" — i.e., the audio content — not silent visual capture. Secondary sources uniformly state that video recorded without an audio track falls outside the statute's all-party-consent requirement, and this reading is consistent with the statutory text and with Washington cases (e.g., State v. Clark, 129 Wn.2d 211, 916 P.2d 384 (1996)) that analyze whether a "conversation" — an inherently aural concept — was private, not whether an image was captured. I could not, however, locate a single Washington appellate decision squarely holding "silent video, standing alone, is categorically outside Chapter 9.73 RCW" — the conclusion is a strong statutory-text inference reinforced by secondary sources rather than a one-case controlling holding, so treat this specific sub-point as medium confidence even though the overall all-party-consent finding is high confidence. Practical implication: a wearable capturing video only (mic off/muted) is very likely unrestricted by RCW 9.73.030; turning on audio brings the recording back under the all-party consent rule for any "private" conversation captured.

What counts as a “private” conversation →

RCW 9.73.030(1)(b) prohibits recording a "[p]rivate conversation, by any device electronic or otherwise designed to record or transmit such conversation... without first obtaining the consent of all the persons engaged in the conversation"; (1)(a) applies the same all-party rule to "[p]rivate communication transmitted by telephone, telegraph, radio, or other device... without first obtaining the consent of all the participants in the communication." (Verbatim text independently re-confirmed against app.leg.wa.gov.) The statute does not define "private" (confirmed: Clark itself says "The Legislature did not define the term 'private' in RCW 9.73"), so courts fill the gap with a reasonable-expectation-of-privacy test rather than a bright-line rule. CORRECTED DOCTRINAL LINEAGE (the submitted draft had this backwards): the test does NOT originate in State v. Christensen (2004) with Clark (1996) "applying" it -- that is chronologically impossible, since Clark was decided eight years before Christensen. Independently reading the full opinions (not just secondary summaries) on CourtListener: the "intent or reasonable expectations of the participants as manifested by the facts and circumstances of each case" formulation traces to Kadoranian v. Bellingham Police Dep't, 119 Wn.2d 178, 190, 829 P.2d 1061 (1992) (quoting State v. Forrester, 21 Wn. App. 855, 861, 587 P.2d 179 (1978), for the dictionary definition of "private"). State v. Clark, 129 Wn.2d 211, 225-27, 916 P.2d 384 (1996), applies and elaborates that Kadoranian test into the operative multi-factor reasonableness analysis -- duration and subject matter of the conversation; location and presence/potential presence of a third party; and the role of the nonconsenting party and his/her relationship to the consenting party -- and on that basis held that drug-sale conversations on public streets, in front of or within earshot of third parties/passersby, were NOT "private," so no one's consent was required to record them. The precise "(1) subjective intention ... (2) reasonable" two-prong phrasing was first articulated in State v. Townsend, 147 Wn.2d 666, 673, 57 P.3d 255 (2002) (which itself cites Clark), then crystallized as an explicit two-part test in State v. Christensen, 153 Wn.2d 186, 193, 102 P.3d 789 (2004) (citing Townsend for the two-part phrasing and Clark, 129 Wn.2d at 225-27, for the multi-factor list), and restated again in State v. Modica, 164 Wn.2d 83, ¶ 7, 186 P.3d 1062 (2008) (citing Christensen). I independently pulled and read the full text of both Christensen and Modica on CourtListener to confirm this citation chain directly (not from secondary paraphrase). State v. Flora, 68 Wn. App. 802, 806-08, 845 P.2d 1355 (Div. I 1992), applies the same Kadoranian/Clark-line gate to hold that an on-duty officer's statements during a public arrest are not a "private conversation" (confirmed: Clark's own opinion cites and summarizes Flora this same way). So Washington does have a real, substantive private/non-private gate -- the corrected point is only about which case is the doctrinal wellspring (Kadoranian/Forrester, elaborated by Clark, later re-phrased by Townsend/Christensen/Modica), not about the existence or substance of the gate itself, which the original analysis got right. A communication must first qualify as a "private conversation"/"private communication" under Washington's subjective-intent-plus-objective-reasonableness test. That test originates in Kadoranian v. Bellingham Police Dep't (1992) (quoting State v. Forrester (1978)) and was elaborated into its operative multi-factor form -- duration/subject matter, location and presence of third parties, and the parties' relationship -- by State v. Clark, 129 Wn.2d 211, 225-27, 916 P.2d 384 (1996); the precise "(1) subjective / (2) reasonable" two-part phrasing was later articulated in State v. Townsend (2002) and restated in State v. Christensen (2004) and State v. Modica (2008), both of which cite back to Clark and Townsend rather than the reverse. If a communication clears that private-conversation threshold, Washington's all-party consent rule (RCW 9.73.030) requires every participant's consent (or the RCW 9.73.030(3) recorded-announcement mechanism) before it may be recorded; if it does not clear the threshold -- e.g., a transaction conducted on a public sidewalk within earshot of passersby (Clark) or an officer's statements made during a public arrest (Flora) -- no consent is required from anyone, regardless of Washington's all-party classification. Net effect on the dataset's consent_type='all_party' classification: unchanged and correct: the gate determines whether the all-party rule is triggered at all, it doesn't change which consent rule applies once triggered.

Biometric / voiceprint law →

Washington Biometric Privacy Act (RCW 19.375) — "Voiceprint" is named in the inclusion list of "biometric identifier" under RCW 19.375.010 ("data generated by automatic measurements of an individual's biological characteristics, such as a fingerprint, voiceprint, eye retinas, irises, or other unique biological patterns..."). HOWEVER — independently confirmed from the statute's own text, and not addressed in the prior draft of this research — the same definition contains an exclusion clause: "'Biometric identifier' does not include a physical or digital photograph, video or audio recording or data generated therefrom, or information collected, used, or stored for health care treatment, payment, or operations under [HIPAA]." (Verbatim, confirmed independently via two official legislature URLs.) Because live.law's proposed feature derives a voiceprint FROM a captured audio recording, there is a genuine textual tension: the voiceprint may be "data generated" from an "audio recording," which the statute expressly excludes from "biometric identifier" — meaning RCW 19.375 may not apply to this feature at all, not merely apply with a narrower duty. No Washington case law addressing this specific inclusion-vs-exclusion tension was found (search budget for this session was exhausted before that could be more thoroughly checked, so absence of a hit is not dispositive). Separately, and assuming for argument's sake the statute is found to apply: RCW 19.375.020(1) requires notice, consent, or an opt-out mechanism before enrolling a biometric identifier "for a commercial purpose" — a term defined narrowly in .010 as enrollment in furtherance of selling/disclosing the identifier to a third party for marketing of goods/services unrelated to the original transaction, and confirmed to exclude security/law-enforcement purposes. Purely internal enrollment to power live.law's own product feature, never sold/disclosed to third parties for unrelated marketing, may fall outside this trigger on a plain reading. Independently of the .020(1) question, if "biometric identifier" status is found to attach at all: (i) .020(3) bars selling/disclosing an enrolled identifier to third parties for a commercial purpose without consent (subject to listed exceptions); (ii) .020(4) requires "reasonable care" against unauthorized access and caps retention to what is necessary for legal compliance, fraud/security prevention, or providing the enrolled-for service; (iii) .020(7) has an untested "security purpose" carve-out from the notice/consent duty; (iv) .020(5) requires new consent before using/disclosing an already-enrolled identifier in a manner materially inconsistent with original terms. Net: there is a real, unresolved argument that RCW 19.375 does not reach a voiceprint derived from recorded audio at all, which is a materially different and more favorable posture for live.law than "voiceprint is squarely covered" as previously stated — but this is an inference from statutory text, not a settled legal conclusion, so it should not be treated as a green light without counsel review.

Attorney ethics (ABA Formal Op. 01-422) →

Independently re-verified and confirmed. I directly queried the live WSBA Advisory Opinions database (ao.wsba.org) myself (not merely reading its search-form description) by submitting the actual ASP.NET search form and reading the results: - Keyword "surreptitious" -> 0 of 1,550 opinions matched. - Keyword "eavesdrop" -> 0 of 1,550 opinions matched. - Keyword "wiretap" -> 0 of 1,550 opinions matched. - Filtered to RPC "4.4" -> exactly 9 opinions (matching the researcher's count), none concerning recording (one is about metadata in electronic documents; the word "record" appears only in unrelated senses). - Filtered to RPC "8.4" -> 84 opinions total (the researcher's write-up said "10+," which is a significant undercount of the actual number returned by that filter, though it doesn't change the substantive finding since the more targeted eavesdrop/wiretap/surreptitious searches returned zero hits across the entire 1,550-opinion database, not just the RPC 8.4 subset). - Keyword "recording" -> 66 of 1,550 matched; the first page of results (visually scanned) contains no opinion on point (hits are false positives like public-records requests, medical records, and real-property recording). I also independently pulled and read the actual text of RPC 8.4 and RPC 4.4 (via the official courts.wa.gov PDFs, converted to text) and confirmed: (1) RPC 8.4(c) and RPC 4.4(a) are quoted accurately, and (2) neither rule's official comments (including the Washington-specific comments) mention recording, consent, surreptitious conduct, or ABA Formal Opinion 01-422/337 anywhere in the text. I confirmed via CourtListener's public search API (official case database) that In re Disciplinary Proceeding Against Huddleston, 137 Wn.2d 560, 974 P.2d 325 (1999), is a disbarment case that turns on "violating the Rules of Professional Conduct" in a matter that also full-text-matches a search for "Huddleston magazine subscription" -- corroborating the researcher's characterization that this is an unrelated fraud/business-scheme case, not a recording case, and that the earlier keyword hit was a false positive on the word "record" (as in "the record does not support the conclusion"). I confirmed RCW 9.73.030's operative text directly from app.leg.wa.gov and it matches the quoted all-party-consent language. I fetched the MyShingle 2026 article and confirmed it exists and states the ABA position is "divided," citing a University of Illinois Chicago Law Review count of "at least eighteen states" finding recording not unethical versus "advisory opinions in at least nine other states" finding it unethical. One caveat: the MyShingle article does not name any specific states in either group (it only cites the aggregate counts), so the researcher's statement that "Washington is not named in either group" is technically accurate but is not itself informative -- no state is named in this source, so it neither confirms nor rules out Washington's placement. This is a minor overstatement of what that particular secondary source shows, though it does not affect the bottom-line conclusion, which rests primarily on the absence of any on-point WSBA opinion or Washington case, independently confirmed above. Net effect of verification: the classification of "no_direct_authority" is correct and, if anything, is now more strongly supported than the original write-up indicated, because I ran the live-database zero-result keyword searches (surreptitious/eavesdrop/wiretap) myself rather than relying on the researcher's unverified description of having done so.

Can you personally sue over a violation?

Yes — this state gives a private right of action, not just criminal/regulatory enforcement.

Can I record as a guest in someone else's home?

It depends on the specifics of the moment — Washington doesn't give a guest the same clear leeway a homeowner gets, so this is closer to a case-by-case call than a fixed yes or no.

What if I'm on a call with someone in another state?

Play it safe: if either state on the call requires everyone's okay to record, treat the whole call that way — don't assume Washington's rule is the only one that matters just because you're the one wearing the glasses.

If it's legal to record, can it still be used against me later?

Usually, yes. Being legal to record and being protected from a future court case are two different questions — a legal recording can generally still be obtained by the other side later. That's especially true the moment a third person is present or on the line — even one extra listener can be enough to strip away legal protection here. Recording your own conversation with your own attorney is a separate, more protected situation, but almost everything else — a landlord, a police officer, a family member — isn't automatically off-limits to opposing counsel just because it was legal to capture.