Synthetic people or voice
Does the work depict a real or invented person using generated image, voice, or likeness?
The Third Asset · Verified 18 August 2026 · Review due 18 September 2026
Move from an argument about AI risk to a decision record your creative, legal, rights, data, and operations teams can actually use.
Interactive workbench
Answer six gating questions. “Unknown” is treated as an unresolved control—not as a safe answer.
Does the work depict a real or invented person using generated image, voice, or likeness?
Could an audience reasonably read the content as a person’s experience, opinion, or recommendation?
Were client, creator, licensed, confidential, or scraped materials supplied to an AI system?
Does the output imitate a recognizable artist, voice, composition, recording, or protected character?
Is there no tested decision for when, where, and how synthetic or sponsored content will be disclosed?
Is there no named owner for complaints, misuse, impersonation, takedowns, or material output changes?
Operating model
Capture use case, people, inputs, vendors, claims, territories, and distribution before production.
Assign who can approve, pause, remediate, document, and release—by risk, not hierarchy alone.
Attach permissions, source records, tests, disclosures, approvals, and vendor terms to the decision.
Watch for drift, misuse, complaints, platform changes, and incidents after the campaign ships.
This tool produces a structured issue-spotting brief. It does not determine legality, clear rights, replace counsel, or guarantee regulatory compliance. Scores are workflow triage—not a validated risk model.
Primary source layer
Links go to the issuing U.S. agencies. Verify current law, jurisdiction, platform rules, and contract terms for the actual campaign.
Truthful endorsements, material connections, virtual influencers, and clear disclosures.
The final rule addresses fake or false reviews and testimonials, including specified AI-generated material.
Primary reports and guidance on digital replicas, copyrightability, and AI training.
Voluntary cross-sector risk-management actions for generative AI.
Connected product layer