1. Purpose
Deepacal provides powerful synthetic media tools. This Policy sets the standards that keep that power lawful and ethical, and reflects emerging obligations including the EU AI Act transparency duties for deepfakes, state-level synthetic media statutes and platform disclosure rules.
2. Consent-first principle
Every likeness and every voice used on Deepacal must be either your own or covered by documented consent from the individual. Consent must be specific, informed, freely given, recorded and revocable. You must be able to produce evidence of consent on request; failure to do so is grounds for suspension.
Consent from an agency, employer or rights-holder does not replace consent from the individual where biometric law requires it.
3. Disclosure and labelling
- Label AI-generated or AI-altered media when publishing it, especially where a viewer could reasonably believe it is authentic.
- Never remove, obscure or falsify provenance metadata, watermarks or disclosures applied by Deepacal or its providers.
- Use platform-native AI disclosure controls on social networks where they exist.
4. High-risk uses requiring extra care or prohibited
- Political, electoral and public-safety communications — prohibited where deceptive; always disclose otherwise.
- Financial promotion, investment advice or endorsements attributed to real individuals — prohibited without documented authorisation.
- Medical, legal, safety or emergency instructions delivered by synthetic presenters — must carry clear human accountability and disclosure.
- Court, insurance, journalistic or evidentiary contexts — synthetic media must never be presented as authentic recording.
- Biometric authentication, identity verification and age verification — strictly prohibited.
5. Human oversight
You must review AI output before publication. Automated pipelines, batch generation and scheduled Studio jobs do not remove your obligation to review each output for accuracy, consent compliance and legality.
6. Safety measures we apply
- Feature gating, credit accounting and rate limiting to reduce automated abuse.
- Account verification, two-factor authentication and privileged-action audit logging.
- Provider-level safety filters, plus removal and suspension on report or detection.
- Retention limits and private, signed-URL storage of generated media.
- Cooperation with lawful requests and with victims of non-consensual synthetic media.
7. Reporting misuse
If you believe your likeness or voice has been used on Deepacal without consent, contact abuse@deepacal.com with evidence of identity. We prioritise these reports, and will remove content, disable the offending account and preserve records for law enforcement where appropriate.