Available now · Local download
AI Content Evidence Report
The current free checker lets you download a structured decision record with the facts you entered, rule version, source references, separate decision layers and unresolved review steps. Keep that JSON yourself: Disclara does not store reports or verify a platform-served asset. Shared reviewer records, client workspaces and campaign exports remain planned, not features of the downloadable report.
What the report is for
Use a downloadable decision record to show a reviewer which facts were supplied, which rule versions were applied, what action was recommended and where uncertainty remains. Keep the file in your own approved storage. Disclara does not currently save it in an account or authenticate the identity of whoever supplied the answers.
Illustrative example · Not a completed check
Sample decision for an agency reviewer
Assumptions: a brand plans to publish a realistic, fully synthetic spokesperson video as a YouTube advertisement in the EU. The character could appear authentic. No actual media file, client policy, rights document or platform upload was inspected. Ruleset shown: 2026.10.03-ph-documentation-dependencies.1.
Next step: disclose and obtain accountable review
On these stated facts, identify the synthetic spokesperson and voice clearly, use YouTube’s applicable altered-content control, and have a responsible person review claims and rights before publication. This is a demonstration of the report format, not approval of a real advertisement.
Why
- EU law: realistic synthetic media that could falsely appear authentic can trigger Article 50(4) disclosure.
- Platform: YouTube’s altered-content process is assessed separately from EU law.
- Client policy: not supplied; permission is not established.
What remains unverified
- Rights, consent and accuracy of product claims.
- Exact final asset, embedded provenance and served variant.
- Whether a native or visible disclosure was actually published.
Evidence to retain: the actual checker JSON, approved asset/version hash, accountable reviewer, publishing settings and a snapshot of the served ad. The example above contains none of those proofs.
Check a similar case with editable assumptions →How to create and use your own record
- Enter the actual asset, AI operations, publication destination, date and responsibility in the checker.
- Read the next action, separate legal and platform layers, and unresolved human-review gates.
- Download the JSON and store it with the approved asset and any later publication evidence.
- Replay the JSON in the browser verifier when you need to check whether its recorded inputs reproduce the decision.
Current limitations
- The free checker does not retain reports, manage reviewers or provide client workspaces.
- Hashes and references supplied by a user do not authenticate the underlying asset or approval.
- Approved-versus-served fields are declarations, not an automatic platform capture.
- Replay checks internal consistency against the current evaluator; it does not verify the author or a trusted timestamp.
Last product description checked: . Compare current and planned workspaces →
Advanced: illustrative technical record fields
This field sketch is explanatory only; it is not an exported, signed or verified decision.
Evidence Record — anatomy
One immutable record per approved and served version: what changed, what was permitted, which rule applied and who approved it.