AI Transparency Notice

JARAI STUDIO Ltd ("JARAI", "we", "us") is a generative-AI media production platform. This notice explains how we meet the transparency duties of Article 50 of the EU AI Act and how we label and trace the content our platform produces. It complements our Terms of Service, Privacy Policy and Acceptable Use Policy.

1. Content on our platform is AI-generated or AI-assisted

JARAI's production pipeline creates media — text, images, audio (including synthetic voice), and video — using artificial-intelligence models. Output you generate, download, or publish through JARAI is, unless you tell us otherwise, AI-generated or AI-assisted content. Where you combine AI output with your own material, the AI-produced portions remain AI-generated.

2. Disclosure obligation (Article 50(2))

Article 50(2) requires that AI-generated or manipulated content be marked as such in a machine-readable format and disclosed to recipients. We split this duty between what the platform applies for you automatically and what remains yours as the publisher:

• What JARAI does. Every account carries an AI-content disclosure policy that is on by default. For productions it covers, JARAI automatically appends a disclosure statement to the platform description and to the production's transcript — this happens before you publish, not left for you to remember afterwards. Accounts can additionally choose a visible on-screen disclosure burned into the video itself for higher-visibility cases.
• Machine-readable provenance. Where our content-provenance manifest embedding is live for a deliverable's format, JARAI embeds a manifest (see §3) recording that it was AI-produced. We are rolling this out across formats and are still confirming complete coverage in production.
• Your disclosure as publisher. Some platforms and jurisdictions require their own specific label or format — for example a platform's native "AI-generated" tag, or a disclosure wording a particular sector or country mandates. Applying those platform- or jurisdiction-specific labels remains your responsibility as the publisher of record (our Terms make you that). JARAI's default disclosure is a baseline, not a substitute for a label a specific platform or law requires you to apply yourself.

3. Content provenance (C2PA)

JARAI embeds a C2PA-aligned provenance manifest in the media it generates. The manifest records the producer (JARAI Studio Ltd), the category of AI model used, and the UTC time of generation, so downstream tools and recipients can confirm the content's origin.

We are rolling out provenance embedding across formats — format-native metadata for supported image formats, and cryptographically-signed manifests for audio and video — and are still confirming complete coverage of this in production. Account holders can verify the provenance of any JARAI-produced deliverable through the platform.

4. Models we use

JARAI orchestrates third-party AI models across the production pipeline, broadly in these categories:

• Large language models — narrative, script, caption, and metadata generation.
• Image-generation models — scene and asset imagery.
• Speech-synthesis models — voiceover and synthetic voice.
• Video-composition models — assembled video deliverables.

Our model providers are listed on our sub-processors page, with what each receives and where. We select and change providers over time to deliver the best quality; provenance manifests record the model category used for each deliverable.

5. Prohibited and high-risk uses

Some uses of AI-generated content are prohibited regardless of disclosure — for example content that is unlawful, or that impersonates a real person without authorisation. These are set out in our Acceptable Use Policy, and a small set of categories are blocked at generation time and cannot be overridden.

6. Reporting and contact

To report content you believe infringes your rights, use our Takedown Procedure. For questions about this notice, contact us.