Approval and rights review

Before any production is published, JARAI shows you a Rights Review screen with a list of potential risks JARAI’s analysis surfaced. You decide whether to publish — JARAI doesn’t block you.

This is by design: JARAI is an informer, not a blocker. Our analysis surfaces things that might matter; you, as the publisher, decide whether to proceed.

🎬Read the Rights Review and decide

Prefer to read? Open the step-by-step transcript

1. Open a production that’s awaiting publish — the Rights Review screen lists any advisory warnings (copyright, likeness, privacy, attribution), per platform.
2. Read each warning — it tells you what JARAI’s analysis flagged and why.
3. Acknowledge the ones you accept; deselect any platform you don’t want to publish to.
4. Approve to publish (with the risks acknowledged on record), or Reject.

Why this matters legally

Under UK and EU law, the publisher of a piece of content bears primary civil liability for what it contains (copyright, right of publicity, defamation). When you click Publish, you confirm that you have the rights to use the content. Our Terms of Service include a clause that shifts this liability to you — that’s the standard model used by Midjourney, Runway, Pika, ElevenLabs, and most other generative AI tools.

What you’ll see

Per platform (LinkedIn, TikTok, etc.), the Rights Review screen lists every warning flag the production carries:

Flag	What it means	What to do
ContentSafetyBelowFloor	Our analyser scored an image below the per-account safety threshold (e.g., possibly suggestive composition, possibly violent visual).	Look at the image. If it’s fine for your audience, proceed. If not, replace the image (regen) or reject the production.
CommercialUseDisallowed	An asset (usually a stock image) was flagged as not licensed for commercial use.	Either find a different image or accept the legal risk of publishing anyway.
CopyrightRiskHigh	An image visually resembles known copyrighted material.	If you have rights to use it (e.g., you own the source), proceed. Otherwise, reject.
AttributionRequired	A stock image needs to credit the photographer.	Click Add attribution to type in the credit text; or proceed acknowledging you’ll add it externally.
IdentityLabelsPresent	The image appears to depict an identifiable person.	If you have a release form, proceed. If it’s a public figure where editorial use applies, proceed. If unsure, reject.
PromptAlignmentLow	The generated image doesn’t closely match the scene’s image prompt.	Often cosmetic; preview it. If it works for the scene, proceed. Otherwise regenerate.
QualityScoreLow	The image’s overall quality score from our analyser is below the threshold.	Usually a soft signal; preview, then decide.

How to decide

1. Click each warning to see the underlying observation values (the actual safety score, the resemblance match, etc.).
2. Tick the “I understand and accept” box per platform you want to publish to. Unticked platforms are skipped.
3. Click “Dispatch to platforms” when you’re ready.

Your acknowledgement is logged with a timestamp. JARAI keeps this log for audit (and as evidence under our ToS that you reviewed the risks).

What JARAI WILL block

A very small set of content categories triggers a hard block that you cannot override:

• CSAM (child sexual abuse material) — detected by an independent classifier
• NCII (non-consensual intimate imagery)
• Terrorist content
• Content targeting specific real children

These categories are illegal under criminal law in essentially every jurisdiction. JARAI’s provider liability for facilitating them is non-delegable.

See also

• Your privacy rights (GDPR) — distinct from rights review; covers personal data
• Submit a takedown notice — if you believe content JARAI produced infringes your rights
• Legal — Terms of Service, Privacy Notice, Acceptable Use Policy