Child Safety
Effective October 19, 2026 | Last updated October 19, 2026
1. Our position #
Kizen has zero tolerance for child sexual abuse and exploitation material. We do not permit it to be uploaded to the Service, generated with the Service, stored on the Service, or shared through it, and we act on every credible report.
This policy covers everything that moves through Kizen: images you upload, prompts you write, text and images the Service generates, and anything shared in a Kizen community or published on a Kizen surface.
2. What is prohibited #
The following are prohibited absolutely, with no exception for artistic intent, fictional framing, satire, or research.
Sexualised depictions of minors. Any image, video, text or audio that depicts a person under 18 in explicit sexual activity, or that depicts the sexual organs or anal region of a person under 18 for a sexual purpose.
This applies whether or not a real child exists. It covers photorealistic generated imagery, stylised and animated output, drawings, and anything created, adapted or modified to appear that an identifiable minor is involved. Canadian law is explicit that a visual representation is covered whether or not it was made by electronic means, and that it extends to imaginary persons. United States law separately covers computer-generated imagery indistinguishable from a minor, imagery modified to depict an identifiable minor, and obscene stylised material where no real minor exists.
Photorealistic depictions of minors in any sexualised context, including suggestive posing, sexualised framing, and content designed to be read as sexual by an ordinary viewer, whether or not it is explicit.
Uploading an image of a minor as a reference. You may not upload a photograph of a person under 18 and use it as an image-to-image reference, a character reference, or any other generation input.
Prompts that combine a minor descriptor with a sexual descriptor, in any phrasing, including attempts to evade filters through indirection, euphemism, coded language, character age assertions, or fictional framing.
Grooming, solicitation and facilitation. Content that sexualises a minor for the purpose of grooming, that solicits sexual contact with a minor, or that provides guidance for abuse.
Ban evasion. Creating a new account after removal for a breach of this policy.
3. How we detect it #
We do not rely only on reports.
On upload. Every image uploaded to the Service is checked against known child sexual abuse material using perceptual hash matching, and is screened by a classifier trained to identify material not previously seen.
Before generation. Prompts are screened before they reach any model. Prompts that combine minor descriptors with sexual descriptors are refused, as are known evasion patterns.
On reference images. Uploads used as generation references are screened for apparent minors, and generation is refused where a minor is detected.
After generation. Generated output is screened by the same classifiers.
Human review. Anything flagged is escalated to a small, named group for review. We deliberately limit who can see flagged material.
We also select and configure our third-party model providers with child safety in mind, and we do not disable or circumvent the safety filtering those providers apply.
Nothing in this section requires or authorises us to seek out this material, and we do not.
4. What happens when we find it #
Immediate removal of the content from the Service.
Permanent account termination, without notice and without an opportunity to cure. No refund of any subscription or credit balance.
Ban evasion monitoring, so that a terminated user cannot simply register again.
Preservation. We preserve the material and the data around it, including account identifiers, network and timestamp data, prompts and generation parameters, in a restricted and access-logged store, for as long as the law requires and no longer.
Reporting. We report to the authorities described in section 5.
For legal reasons, we may not be able to tell a user why their account was terminated, or confirm whether a report has been made.
5. Who we report to #
Canada. Studio Kizen Corp. is an Ontario company and reports to the Canadian Centre for Child Protection through Cybertip.ca, as required by Canadian law. Where we have reasonable grounds to believe our Service has been used to commit an offence, we notify law enforcement.
United States. We are registered with the National Center for Missing and Exploited Children as an Electronic Service Provider and report to the NCMEC CyberTipline.
We report material whether it was uploaded to the Service or generated through it.
6. Reporting to us #
If you find content on Kizen that breaches this policy, tell us immediately at hello@kizenworlds.com. Put "Child Safety" in the subject line so it is escalated.
You do not need a Kizen account to report. Include a link or description of where you saw the content. Do not attach or forward the material itself.
We treat these reports as our highest priority.
If you are aware of a child in immediate danger, contact your local emergency services first.
7. Age of use #
Kizen accounts, including paid features, are for people aged 18 and over. Accounts we find to belong to anyone under 18 are closed and the associated data deleted.
Kizen is not directed to children. Every title published on Kizen is reviewed by a person and assessed against published criteria for whether it is directed to children, and work that is directed to children is not published.
8. How we build #
We align our approach to the Safety by Design for Generative AI principles developed by Thorn and All Tech Is Human, which structure child safety commitments across how a system is developed, deployed and maintained.
We keep this policy under review and update it as our detection and reporting practices change.
9. Contact #
Studio Kizen Corp. 350 Bay St, Suite 1300B Toronto, Ontario M5H 2S6, Canada