AIAI Video Lab

SAFETY & PRODUCT CONTROLS

Controls you can see.
Decisions people still own.

AI Video Lab keeps review, spending, revision, and export decisions visible. This page separates the controls that exist today from the responsibilities that remain with creators and their teams.

CURRENT PRODUCT CONTROLS

Trust should be visible in the workflow.

These controls describe the current AI Video Lab implementation. They do not claim model-level filtering, automated moderation, or certifications that the product does not have.

WORKSPACE CONTROLIMPLEMENTED

A stage cannot approve itself.

Story, scenes, assets, audio, and later production stages carry a visible review state.

CONTROL PATH
REQUESTDraft
VISIBLE CHECKPOINTReviewer decisionHuman decision
HOLDReturn with notes
ADVANCEUnlock next stage
  1. 01Draft
  2. 02Submit
  3. 03Human review
  4. 04Approve or revise
What it blocks
Automation pauses when the next required stage has not been approved.
What remains visible
Submission status, reviewer or approver identity, timestamps, and requested changes remain with the production record.
What a person decides
A permitted workspace member chooses whether the stage is approved or returned for changes.

THE RESPONSIBILITY BOUNDARY

Clear about what the product does and does not do.

01

Editorial state is explicit

Stages can be drafted, submitted, approved, or returned for changes. Required approval stops the automated path.

The product does not replace the contextual judgment of a creator, reviewer, parent, educator, or studio.

02

Paid actions are gated

Paid generation defaults off and requires server enablement, owner confirmation, and a budget check.

The product cannot guarantee that a connected provider will remain available, produce suitable output, or keep the same pricing.

03

Publication stays separate

A technical render must still pass production checks and be deliberately selected for export.

A completed export is not legal, editorial, developmental, or safety approval.

04

Claims stay narrow

This page documents controls that exist in the current product code and production workflow.

AI Video Lab does not currently claim automated moderation, model-level filtering, C2PA, compliance certification, or independent safety auditing.

05

Provider access stays user-scoped

Connected API credentials are encrypted before storage, selected by authenticated user ID, shown only as masked hints, and removable from account settings.

Connecting a provider authorizes eligible server-side requests. Provider access, terms, pricing, rotation, and revocation still belong to that provider account.

REPORT A CONCERN

Give us enough context to investigate.

Include the project or page, the production stage, what you expected, and what concerned you. Do not email passwords, provider keys, private identity documents, or sensitive information about a child.

Email a safety concern Read the privacy policy

BUILD WITH DELIBERATE CHECKPOINTS

Plan the work, review each stage, and keep the final decision human.

Open production guides See the complete workflow