Archived copy. This is Consumer Health Data Privacy Policy, Version 1.0 — Effective 20 October 2026, exactly as Kavi published it. This page is never edited. SHA-256 of the source text: 9dfc31893a3c7ce5e0b688d94ec0062c639dab300711d0917fa4d171df1bb8d8. Current version: kavi.audio/legal/consumer-health-data-privacy-policy. All versions: kavi.audio/legal/policies.json.

Consumer Health Data Privacy Policy

Kavi Media, Inc.
Version 1.0 — Effective 20 October 2026
Published at kavi.audio/legal/consumer-health-data-privacy-policy

1. About this Policy

1.1 This Consumer Health Data Privacy Policy is published under Washington's My Health My Data Act, chapter 19.373 RCW (the "Act"). It covers consumer health data, as the Act defines it, about consumers as the Act defines them: people who live in Washington State, USA, and people whose consumer health data is collected in Washington.

1.2 Kavi Media, Inc. is a Delaware corporation with its principal place of business in Alpharetta, Georgia, USA (mailing address: 169 Madison Ave STE 30314, New York, NY 10016, USA) ("Kavi", "we", "us"). Kavi decides how the consumer health data described here is collected, used and shared.

2. The consumer health data Kavi collects

2.1 Kavi collects only these categories of consumer health data:

Category Whose What it includes
Voice recordings Creators (including co-hosts, co-owners, short-form creators, musicians and authors), guests, cast members, narrators, and people who apply to join Kavi as creators Recordings of your voice in content delivered to or made for Kavi: episodes, shows, music releases, artist spoken intros, audiobooks, readings, trailers, clips and Bites (short-form audio); versions of that content that still contain your voice (for example, a subtitled version); raw recordings, takes, pickups and stems sent for editing; the recorded verbal consent file; and audio samples you link to when you apply to join Kavi. Kavi treats every recording of a person's voice as biometric data under the Act, because an identifier template could be extracted from it, whether or not one ever is.
Health information in your recordings and their transcripts The same people Anything you say in a recording about your own physical or mental health, and the transcript of it.
Voice Data, Voice Models and Replica Outputs Only a creator who signs an AI-Voice & Digital Replica Consent and ticks its biometric consent, or an adult voice actor who signs a Voice License Agreement and gives its Washington consents. Never a guest, cast member or narrator in that role, and never anyone under the age of majority where they live. The recordings of your voice used to make your voice model ("Voice Data"), the model itself ("Voice Model"), and audio made with it in your voice ("Replica Outputs"), as that Consent defines them.

2.2 Kavi does not collect any other consumer health data. Kavi does not collect consumer health data from listeners: Kavi does not use what listeners play, like, share, follow or search for, or the interests they choose, to identify or infer anyone's health.

3. When Kavi collects it

3.1 Creators. Kavi collects and shares your voice recordings and the health information in them only as needed to provide the service you asked for under your agreement with Kavi (receiving, reviewing, distributing, promoting and paying for your content). If you apply to join Kavi, Kavi listens to the audio samples you link to only to decide whether to sign you.

3.2 Guests, cast members and narrators. Kavi collects and shares your voice recordings, and the health information in them, only with the two separate Washington consents at your signature: one to collection, and a separate one to sharing and publication. They are in the Guest / Participant Release (clause 9.2), the Cast & Writer Release (clause C2), the Narrator & Translator Services and Assignment Agreement (at its signature blocks) and, for recordings made in Kavi Studio online, the join-screen consent (Studio Use Agreement and Liability Waiver, Annex P-2). A recorded verbal consent does not give these consents. Kavi's process requires both consents before it accepts a recording; if a recording reaches Kavi without them, Kavi does not use it.

3.3 Training Kavi's own AI models. Kavi uses voice recordings for this (section 4.1(c)) only if you ticked the separate Kavi model training consent when you signed. Kavi never does this for anyone under 18 (or under the age of majority where they live, if higher), and never with music.

3.4 Voice Data, Voice Models and Replica Outputs. Kavi collects and uses these only if you sign an AI-Voice & Digital Replica Consent and tick and initial its biometric consent (Schedule J-1, Part 2), and then only for the uses you tick and initial in Schedule J-1, Part 1, or if you sign a Voice License Agreement and give its biometric consent (Schedule O-2) and both Washington consents (Schedule O-3), and then only for the uses you initial in its Schedule O-1, Part B.

4. How Kavi uses it

4.1 Voice recordings and the health information in them. Kavi uses them only for the following purposes, and only as far as your agreement or release allows:

(a) Producing and publishing the content. Receiving and reviewing it; editing it (for example cutting, shortening, reordering, adjusting audio quality, and adding music, intros, credits, sponsorship messages and required disclosures); making clips, trailers and Bites; making translated and dubbed versions, which are voiced by a different human, or by a stock synthetic voice or a Kavi Voice licensed from a consenting voice actor, neither of which is modeled on, matched to or readily identifiable as your voice, and are labeled as translated with a synthetic voice; publishing, distributing and promoting the content; crediting you; and, for creators, paying for it under your agreement.

(b) Running AI tools on the content. Making transcripts, and using them and the audio for captions, search, keyword tags and episode descriptions for the catalog, recommendations of content, chaptering, choosing clips, content moderation (including a private safety and legal-risk note that an AI tool writes from the transcript for the Kavi staff member who approves an episode), fraud detection, and detecting and labeling synthetic audio. No one may apply a voice-cloning, voice-matching or voice-preservation feature of any tool to your voice, except to a creator's own voice under that creator's AI-Voice & Digital Replica Consent, or to a voice actor's own voice in recordings made under that actor's Voice License Agreement.

(c) Training Kavi's own AI models, only with the consent in section 3.3. Kavi uses the audio (such as tone, rhythm and intonation) and the transcript to train, fine-tune, evaluate and improve AI models that Kavi owns or exclusively controls, including speech-recognition, transcription, translation and speech-synthesis (text-to-speech) models. No such model is ever used to imitate or reproduce your voice, and Kavi uses reasonable measures to stop that happening. Kavi may license those models (speech-synthesis models only as a hosted service, never as a copy of the model), but never your recordings or any dataset containing them.

(d) Recording consent and handling requests. Keeping the recorded verbal consent file as proof of consent; and handling your privacy requests, withdrawal and removal requests, complaints, copyright notices and legal orders about the content.

(e) Protecting the service. Preventing, detecting and responding to security incidents, fraud, harassment and illegal activity.

4.2 Voice Data, Voice Models and Replica Outputs. Kavi uses them only to create and run your Voice Model and make Replica Outputs for the uses you ticked and initialed in Schedule J-1, as clause 3 of your AI-Voice & Digital Replica Consent allows, or, under a Voice License Agreement, for the uses you initialed in its Schedule O-1, Part B. Kavi does not merge your Voice Model into any multi-speaker or general-purpose model, and does not use your Voice Data or Voice Model to train any model for anyone else. Kavi's own model training (section 4.1(c)) never uses your Voice Data or Voice Model.

5. Where Kavi gets it

5.1 Kavi collects consumer health data from these categories of sources:

(a) You: content you upload to or deliver through the Kavi Creator Portal or send to Kavi; recordings made at Kavi's recording studio or in Kavi Studio online; Voice Data you supply under an AI-Voice & Digital Replica Consent or a Voice License Agreement; and audio samples you link to when you apply.

(b) The creator of the show you appear in, who records you and uploads the recording to Kavi with your signed release.

(c) Kavi's own systems and the service providers in section 6.2, which make transcripts, edited, translated and dubbed versions, Voice Models and Replica Outputs from those recordings.

6. Who Kavi shares it with

6.1 Third parties. Kavi shares consumer health data only with these categories of third parties, and only as section 3 allows:

Category What is shared, and with whose data
Listeners and the public The published content and its clips, trailers and Bites, on the Kavi app and website, and in promotion as your agreement or release allows (including on Kavi's own accounts on social media and other websites). For Replica Outputs: on the Kavi service only, except that a self-narrated audiobook under Category (iii) of the AI-Voice & Digital Replica Consent may also appear in promotional excerpts of up to 90 seconds on Kavi's own unpaid accounts on other platforms (never in paid advertising), and is distributed to other platforms only if you ticked third-party distribution.
Licensees, distributors and other platforms Only where your signed agreement, release or consent covers it. Creators and narrators: as your agreement allows; for example, for an audiobook, the author or publisher whose work is narrated, and platforms the audiobook is licensed or distributed to under the Audiobook Master Agreement or your Narrator & Translator Services and Assignment Agreement. Guests and cast members: only under your separate Washington sharing consent (Guest / Participant Release clause 9.2(b); Cast & Writer Release clause C2(b)), which covers Kavi's licensees and distributors.
The creator of the show and its co-hosts and co-owners Guests' and cast members' recordings in the show they appear in, including edited versions of them that Kavi gives back to the creator, and creators' recordings with the co-hosts and co-owners who run the show with them.
Courts, regulators, law enforcement and government agencies Only what the law requires, or what is needed to respond to a valid legal order.
A buyer or successor of Kavi's business Only as your agreement or release allows, and only to a buyer or successor that is bound by the Act.

6.2 Service providers (processors). These process consumer health data only for Kavi, on Kavi's written instructions, consistent with this Policy:

Category What they do
Amazon Web Services Storage, hosting, transcription (Amazon Transcribe) and delivery of audio to listeners
Cloudflare Network delivery and security
Metered (TURN relay) Relays encrypted live-call audio in Kavi Studio online when a direct connection fails; it cannot read or store it
AI tool providers (currently Anthropic, or OpenAI if Kavi switches) Receive transcripts (not audio) to produce keyword tags, descriptions and the safety and legal-risk review
AI voice vendors on Kavi's Approved AI Voice Vendor List (kavi.audio/legal/ai-voice-vendors) Dubbing with stock synthetic voices or a Kavi Voice; Voice Models and Replica Outputs only under an AI-Voice & Digital Replica Consent or a Voice License Agreement
Editors, translators and narrators Kavi engages Editing, translation and narration, under written confidentiality and data-processing terms
Vendors that train Kavi's own models Only with the consent in section 3.3, and solely for Kavi
Kavi's professional advisers (such as lawyers and accountants) Advice, under confidentiality

Each AI vendor is bound by its written terms, together with the opt-out Kavi has exercised under them, not to train, fine-tune, improve or evaluate any model on your recordings, other than Kavi's own models trained solely for Kavi.

6.3 Affiliates. Kavi does not share consumer health data with any affiliate.

6.4 No sale. Kavi does not sell consumer health data.

7. Your rights and how to use them

7.1 Your rights. You have the right to:

(a) confirm whether Kavi is collecting, sharing or selling your consumer health data, and access it, including a list of all third parties and affiliates with which Kavi has shared or sold it, with an active email address or other online way to contact each of them;

(b) withdraw your consent to Kavi's collection and sharing of your consumer health data; and

(c) have Kavi delete your consumer health data.

7.2 How to make a request. At any time:

Creators may also withdraw a consent through the Kavi Creator Portal. You do not need a Kavi account to make a request.

7.3 Confirming it is you. Kavi confirms your identity using information it already holds: for creators, by your signing in to the Creator Portal or confirming the request from the email address in your Portal account; for guests, cast members and narrators, by confirming the request from the email address or phone number in your signed release or agreement; and for people who applied to join Kavi, by confirming the request from the email address you applied with. Kavi never asks for a government identification number or Aadhaar number for this, and never asks you to send identity documents by email. If Kavi cannot confirm your identity with commercially reasonable efforts, it will ask you for more information and will act once it can.

7.4 Someone acting for you. Someone you authorize may make a request for you. Kavi may ask for proof of their authority and may confirm the request directly with you.

7.5 When Kavi responds. Kavi will act on your request without undue delay and within 45 days after receiving it. If it is reasonably necessary, Kavi may extend this once by up to 45 more days; Kavi will tell you of the extension, with the reasons, within the first 45 days. The time Kavi needs to confirm your identity does not extend these periods.

7.6 What happens when you withdraw consent.

(a) Washington collection and sharing consents (guests, cast members, narrators). If you withdraw your sharing consent, Kavi stops sharing your recordings: it removes your recording (the content, or the part with your voice) from the Kavi service and Kavi's promotion, shares it with no one new, and tells any licensee or distributor it shared the recording with to stop using it. If you withdraw your collection consent, Kavi also stops collecting your recordings, which under the Act includes keeping them, so Kavi deletes them as section 7.7 describes. Withdrawal does not cancel the copyright, performer's-right and moral-right grants and waivers in your release or agreement, but Kavi will not use them to share your recording after you withdraw.

(b) Kavi model training consent. Within 30 days after receiving your withdrawal, Kavi stops using your recordings in new training and deletes them from its training datasets. Models already trained are not retrained, and are never used to imitate your voice.

(c) Biometric consent under an AI-Voice & Digital Replica Consent. Your withdrawal is a revocation under clause 9 of that Consent. If you live in Washington, Kavi stops creating Replica Outputs when it receives your notice, and the Deletion Period for your Voice Data and Voice Model runs from the date of your notice, as clause 9.1 of the AI-Voice & Digital Replica Consent provides. Replica Outputs already published stay available as clause 9.2 of that Consent provides, unless you ask Kavi to delete them under section 7.7. A withdrawal under a Voice License Agreement works the same way under its clauses 15, 16 and 26.1: Kavi stops creating Outputs when it receives your notice, deletes your Voice Data and Voice Model, and removes the Outputs you identify if you ask Kavi to delete them under section 7.7.

Withdrawal does not affect anything Kavi lawfully did before it.

7.7 What happens when you ask Kavi to delete. Kavi deletes your consumer health data from all its systems, including archived and backup systems. A copy in a backup system is deleted no later than six months after Kavi confirms receipt of your request (for Voice Data and Voice Models, as clause 9.3 of the AI-Voice & Digital Replica Consent or clause 26.1(c) of the Voice License Agreement provides). For published content, this means Kavi removes your recording (the content, or the part with your voice) from the Kavi service and Kavi's promotion. Kavi's agreements with creators and narrators let it do this whatever their license term or exclusivity says (Creator Master Agreement clause 16.10(e), Music Master Agreement clause 15.2C, Audiobook Master Agreement clause 3.3B, Kavi Creator Terms — Short-form section 13.1A, and Narrator & Translator Services and Assignment Agreement clause 10.1A). Kavi also deletes your recordings from its AI training datasets; models already trained are not retrained (section 7.6(b)). Kavi also tells every service provider and third party with which it shared the data about your request, so that they delete it too. Copies or clips already distributed outside Kavi's control may not be recallable. Kavi keeps only what the Act allows it to keep to prevent, detect or respond to security incidents, fraud, harassment or illegal activity, and a record of your request and Kavi's response.

7.8 Cost. Kavi does not charge for a request. Kavi may decline a request that is manifestly unfounded, excessive or repetitive, as the Act allows, and will tell you why.

7.9 Appeals. If Kavi declines to act on your request, in whole or in part, you may appeal by replying to Kavi's decision, or by emailing privacy@kavi.audio with the subject line "Appeal". Kavi will reply in writing within 45 days after receiving your appeal, saying what it did or did not do and why. If Kavi denies your appeal, you may complain to the Washington State Attorney General:

7.10 No discrimination. Kavi will not discriminate against you for exercising these rights.

8. Contact