AI Content & Voice Policy

1. Synthetic-by-design

CineVoice produces audio using machine-learning models. Outputs are synthetic even when a “voice style” evokes a genre or archetype. No human performer records your specific line in real time; instead, statistical models predict waveforms. You must label or contextualize outputs truthfully when publishing (e.g. “AI voiceover”) if your jurisdiction, platform terms, or industry code requires disclosure.

2. Prohibited impersonation and deceptive uses

You must not create content that a reasonable listener would believe is an actual individual speaking, unless you hold verifiable rights and consents covering voice likeness and any implied endorsement. This includes celebrities, politicians, journalists, corporate officers, and private figures. Parody is not a shield if the primary purpose is deception, harassment, or commercial misappropriation.

Examples of violations: fake emergency calls, forged voicemail instructions to employees, synthetic interviews presented as authentic news, or customer-support scams.

3. Fraud, scams, and social engineering

Any attempt to obtain money, cryptocurrency, credentials, or personal data through misrepresentation is forbidden. We cooperate with payment partners and law enforcement where legally permitted.

4. Political and civic integrity

Do not use CineVoice to create or amplify deceptive political content, including fake endorsements, false statements attributed to officials, or voter-suppression audio. Regional election laws may impose additional duties; compliance is your obligation.

5. Harmful or hateful content

Content promoting violence, terrorism, child exploitation, or severe harassment is banned. We may use hash-matching, keyword lists, and user reports to enforce this policy.

6. Intellectual property and training data

You may not attempt to extract underlying model weights, training corpora, or proprietary prompts from our vendors. You may not use outputs to train competing models if our upstream agreements forbid it — check your enterprise contract if applicable.

7. Moderation workflow

Violations may result in warnings, rate limits, wallet freezes, account suspension, or termination. We strive for proportionality and human review for high-impact decisions when feasible, but automated blocking may occur in emergencies (e.g. CSAM hash hits).

8. User remedies

If you believe your content was flagged in error, contact abuse support with generation IDs and legitimate use justification. Repeated unfounded appeals may be deprioritized.