AI Transparency Guide
Effective Date: September 16, 2025
Entity: 1534583 B.C. LTD. d/b/a "JVO"
Contact: hi@jvo.io
This AI Transparency Guide ("Guide") explains when and how to disclose AI-generated or materially edited media made with JVO. It complements the Terms, Privacy Policy, and AUP. It is designed to help you comply with common legal and platform requirements. This Guide is not legal advice; always check local laws and the rules of the platforms where you publish.
1) When to Disclose
Disclose at first exposure and persistently if the content:
Is wholly or partly AI-generated (images, lip-sync, voice cloning, face swaps, reenactments).
Materially edits reality such that a reasonable viewer could be misled (e.g., changing words spoken, adding a person, altering events).
Depicts any identifiable person via synthetic face/voice—even with consent.
Is used in news, political, health, financial, or other public-interest contexts.
Is sponsored content/ads or otherwise promoted—follow both ad disclosure and AI disclosure norms.
Important: Disclosures do not cure illegality. If content violates the AUP (e.g., non-consensual intimate imagery), disclosure will not make it permissible.
2) What to Say (Plain-Language Labels)
Use simple, unambiguous language. Examples:
On-screen badge (first frame & persistently visible):
"AI-Generated" or "Contains AI-Generated Content"
"Lip-Sync Performed with JVO"
"Synthetic Voice — Actor Consent on File"
Audio disclosure (first 3–5 seconds for audio-only):
"This is an AI-generated voice performance created with consent."
Caption/Description:
"This video uses AI lip-sync to match new narration to original footage."
"Face/voice recreation performed with permission. Synthetic media disclosure."
Hashtags (if platform-appropriate):
#AIGenerated #SyntheticMedia #AILipSync #Disclosure
Keep wording consistent across placements.
3) Where to Put It (Placement & Persistence)
Video: Badge in the first frame and persistent corner watermark or intermittent lower-third every 10–20 seconds; matching note in the description.
Shorts/Reels/Stories: Badge in the first card/segment and in the caption.
Audio/Podcasts: Spoken disclosure at the beginning and in show notes.
Images: Visible overlay/badge and disclosure in alt text/caption.
Ads/Political Content: Follow platform and jurisdiction-specific rules (longer, more prominent labels may be required).
4) Technical Signals (If Available)
Where feasible, preserve or add machine-readable indicators:
Metadata fields noting "AI-Generated" and transformation details.
Watermarks or content credentials (e.g., provenance/manifest approaches where supported).
Do not remove JVO-applied labels, watermarks, or metadata. Doing so may violate the AUP.
5) Consent & Documentation Checklist
Before publishing synthetic media of a real person:
Obtain written, informed consent that covers: the person's name, the specific project/use, the type of transformation (lip-sync, voice cloning, face swap), distribution channels, duration, and revocation terms.
If the person is a minor, obtain verifiable parent/guardian consent; no sexualized content involving minors is allowed.
Keep date-stamped records (PDF or signed e-signature).
Keep source files and a simple audit trail (who uploaded, who approved, when published).
Honor revocation or takedown requests where required by law or contract.
If content could influence health, finance, or elections, seek legal review and add extra-prominent disclosures.
Sample Consent Language (template snippet):
"I, [Name], consent to the creation and distribution of AI-generated media using my likeness and/or voice for [Project/Use], including [lip-sync/voice cloning/face recreation], to be published on [Channels] for [Duration]. I understand the content will be disclosed as AI-generated. I grant the necessary rights to [Your Organization], and I may revoke prospectively as permitted by law."
6) Labeling Recipes (Copy-Paste Examples)
A. 30–60s Social Video (lip-sync)
First frame badge: "AI-Generated Lip-Sync"
Lower-third (0:00–0:05): "Performed with JVO (consent on file)"
Caption: "We used AI lip-sync (with consent) to dub this message. #AIGenerated #AILipSync #SyntheticMedia"
B. Voice-Over Replacement (podcast clip)
Spoken intro (first 5s): "This segment uses an AI voice performance created with consent."
Show notes: "Contains AI-generated voice; speaker approval on file."
C. Face Swap for Parody
On-screen: "Parody — AI Face Recreation"
Description: "Satire using synthetic face recreation. Not a real event. #Satire #AIGenerated"
Note: Confirm satire is lawful in your jurisdiction; avoid defamation and pornographic depictions.
7) High-Risk Contexts (Extra Care)
Health/Legal/Finance: Add explicit disclaimers ("educational only; not professional advice"), provide links to authoritative sources, and avoid portraying outputs as expert opinions.
News/Politics: Label reenactments; avoid depicting false events as real. Consider visible context cards stating what was edited.
8) Common Mistakes to Avoid
Tiny or transient labels that viewers miss.
Disclosures only in the description with none in-video.
Publishing synthetic depictions of someone without consent.
Using disclaimers to justify content that's prohibited by the AUP.
Removing watermarks/metadata or encouraging others to do so.
9) Governance Tips for Teams
Create an internal approval checklist for any synthetic media leaving your organization.
Use two-person review for public-interest or ad content.
Store consent and disclosure screenshots in a project folder.
Keep a register of platform policies (they change often).
Train staff and contractors; make this Guide part of onboarding.
10) Getting Help
Questions about disclosures, consent, or edge cases? Contact hi@jvo.io. For potential violations (yours or others'), see the AUP §6 (Reporting, Moderation & Enforcement).