AI content policy

Never allowed

We block these, and we act on accounts that try them:

• sexual content involving minors, or any sexualised image of a minor;
• intimate or sexual images of a real person, including ones made without their consent;
• deepfakes of a real person made without their consent, and deceptive depictions of real people;
• fake identity cards, documents or currency;
• impersonating a government official, police, a bank or a courier, including voice calls that ask for an OTP, KYC details or money;
• election content in a real candidate's voice;
• fake reviews or testimonials in ads;
• extremist propaganda and instructions for self harm.

Content involving the sexual abuse of children is preserved as evidence and reported to the authorities. It is never shared again.

Real faces need consent

My Models lets you save a face for use across the studios. You can add your own face after a short live selfie check. To add another person, they get a link and complete the live check and consent on their own phone. Public figures and children cannot be added. ReCreate and Avatar Video follow the same rule: no celebrities, and Avatar Video does not allow expert personas such as doctors or lawyers giving advice.

You may upload a photo of a real person only if you are that person or have their written consent.

Voices need consent

Voice cloning is for your own voice. You must be 18 or older, and you record a short consent sentence live, which we keep as proof. Public figures cannot be cloned. You can delete your voice any time, and it is removed everywhere within 24 hours.

When you dub a video and keep each speaker's voice, you confirm that you own the video and that every person speaking agreed.

Age

Chat, My Models, voice cloning and face uploads are for people aged 18 and over. Faces of minors are never shown in public on Kritiverse.

Labels and metadata

Images and videos made on Kritiverse carry a visible "Made with AI · Kritiverse" label, and shared pages show it too. Dubbed videos show an "AI dubbed" label at the start. Files also carry embedded metadata (C2PA where possible, and IPTC tags on images) that marks them as AI generated. Our tools never strip it.

• Do not remove, crop out or hide the label or the metadata.
• When you post, turn on the platform's AI disclosure, such as the altered or synthetic content setting in YouTube Studio.
• Ads made in UGC Ads have an AI label switched on by default. Under ASCI guidelines, a label does not make it acceptable to use a real person's face or voice without consent, or to invent a customer review.
• Product images exported for marketplaces that ask for it get a synthetic performer tag where an AI person appears.

How we check

Prompts, uploads and outputs are checked before you see the result, in English, Hindi, Hinglish and Urdu. If our safety checks are not working, jobs wait rather than run unchecked. If a job is blocked, its credits come back to you. Each blocked attempt counts as a strike, and three strikes suspend an account.

Sharing, Explore and Remix

Everything you make is private by default. When you share to Explore, others can Remix your creation with their own face, product and words, and you are credited automatically. You are responsible for what you share.

Reports and takedowns

Every public item has a Report button. You can also write to our grievance officer.

• Intimate images and impersonation: removed within 2 hours of a valid report.
• Court or government orders: acted on within 3 hours.
• Other reports: acknowledged within 24 hours.

When we act, we block the file, remove the public copy, keep evidence as the law requires, and may suspend the account. Each output has an ID in its metadata, so we can trace it to the job that made it.