IISHKI. Legal information
AI Content Policy
Prohibited content categories, moderation process, reporting channel, and enforcement actions for image, video, audio, and text generation.
1. User Responsibility
By uploading material or submitting a prompt, the User confirms that they have all required rights, consents, and a lawful purpose. Each output must be reviewed before publication: AI may make mistakes, create similarities to real people, or reproduce unwanted elements.
This Policy applies to every feature of the Service that creates or alters material: image generation, video generation, speech and audio synthesis, and text generation.
2. Content Standards and Prohibited Categories
The Service expressly prohibits creating, uploading, storing, and distributing material in the following six categories.
- Sexual and NSFW content. Pornography, sexually explicit images and video, sexualized depictions of real people without their unambiguous consent, intimate images produced by “undressing” or similar processing, and the use of such material for blackmail.
- Violence and gore. Realistic depictions of killing, mutilation, torture, and cruelty to people or animals, glorification of cruelty, targeted threats, and instructions for causing serious physical harm or for producing weapons or malware.
- Hate speech. Material that degrades or incites discrimination, hostility, or violence against individuals or groups based on race, nationality, ethnic origin, language, religion, sex, gender identity, sexual orientation, age, disability, or another protected characteristic, including extremist symbols and slogans.
- Child-unsafe content (CSAM). Any sexual material involving or depicting minors, sexualization of children, grooming, and any attempt to create, alter, or obtain such material, including fully synthetic material.
- Deepfakes and impersonation. Realistic imitation of a specific person’s appearance, voice, or other biometric trait for fraud, deception, political manipulation, identity impersonation, or circumvention of identity checks, and presenting synthetic material as an authentic record of an event.
- Copyright and trademark infringement. Reproduction of protected works, characters, logos, and trade dress without the rightsholder’s permission, circumvention of technical protection measures, and infringement of image and likeness rights, privacy, or third-party contractual restrictions.
Also prohibited: harassment and doxxing; processing another person’s medical, financial, biometric, or other sensitive data without a lawful basis; material encouraging suicide, self-harm, or dangerous conduct; spam, misleading advertising, fabricated reviews and documents; sale of prohibited substances and other unlawful activity.
2.1. Zero Tolerance
The material listed below is removed immediately, the account is banned without prior warning, and information is provided to the competent authorities where required or permitted by applicable law:
- child sexual abuse material (CSAM) and any sexual content depicting minors;
- material promoting terrorism, extremism, or mass violence;
- instructions for producing weapons of mass destruction;
- incitement to genocide or ethnic hatred;
- non-consensual sexual deepfakes of real individuals.
3. Likenesses, Voices, and Synthetic Media
The User may provide personal data of third parties through the Service, including their images, video and audio recordings, and voice, only with the consent of the relevant third party or another legal basis required for transferring and processing such data through the Service. The User must keep evidence of such consent or other legal basis and provide it to the Operator upon request.
For public or commercial use of realistic synthetic material, the User must clearly disclose its AI origin where the context could mislead the audience or disclosure is required by law.
4. Platform Content Moderation
4.1. Automated Screening
Both the User’s request and the generated output are screened. Prompts and uploaded files pass through the safety systems of the model providers (including OpenAI, Google, BytePlus / ByteDance, xAI, Kling, MiniMax, and ElevenLabs) before generation starts, and the output passes the same screening before it is delivered. A rejected request is recorded in the Service with a content-policy flag, no output is delivered, and tokens for that attempt are not refunded. The Service additionally enforces its own limits on request length and composition and keeps a log of every request with the account and model.
4.2. Human Review
Material flagged by automated screening or received through reports enters a manual review queue. Review is performed by the Operator’s responsible personnel, who have access to the Service’s administration panel and are bound by a confidentiality obligation. The public gallery contains only works the User has explicitly published, and any such work may be unpublished following review.
4.3. Periodic Audits
At least once a month, a random sample of published works and rejected requests is reviewed, together with an assessment of how accurately the automated filters performed. Rules and screening parameters are adjusted based on the audit results.
4.4. Severity Classification
| Level | Scope | Review | Timeline and action |
|---|---|---|---|
| L1 — critical | Material listed in section 2.1: CSAM, terrorism, weapons of mass destruction, incitement to genocide, sexual deepfakes of real individuals | Automatic block with confirmation by responsible personnel | Within 24 hours: removal, permanent ban, referral to the competent authorities |
| L2 — high | Serious breaches of the categories in section 2: non-consensual sexualization, deceptive deepfakes, hate speech, realistic violence | Automatic flag with priority manual review | Within 3 business days: removal of the material and account suspension |
| L3 — medium | Other breaches of section 2, including copyright and trademark infringement, doxxing, and misleading advertising | Manual review | Within 5 business days: warning and removal of the material |
| L4 — low | Minor or disputed breaches, spam, missing AI-origin disclosure | Review on report | Within 5 business days: warning and a request to correct the material |
5. Reporting a Violation
Anyone may report unlawful material, whether or not they hold an account. Two channels are available, both monitored by the Operator:
- email support@iishki.com;
- Telegram @maksim_and_ai_support.
Include a link to the material, a description of the violation, and, if the report concerns your own rights, evidence of those rights. Reports about L1 material are handled out of turn; marking the subject line “urgent” is sufficient. All other reports are answered within five business days.
6. Enforcement Actions
The action is chosen according to the severity of the violation and whether it is repeated. The Operator applies only the actions listed below.
| Action | Measure | Applies to |
|---|---|---|
| 1. Content warning | Written notice requiring deletion or correction of the material | First minor violation |
| 2. Content removal | Unpublishing the work and deleting it from the Service | Confirmed L3–L4 violations |
| 3. Feature restriction | Reduced request limits or removal of access to specific models | Repeat minor violations or a single L3 violation |
| 4. Account suspension | Temporary hold on access pending review | An L2 violation or accumulated violations |
| 5. Permanent ban | Termination of access to all features of the Service | An L1 violation or repeated L2 violations |
| 6. Referral to authorities | Retention of the minimum necessary evidence and cooperation with authorities | Suspected unlawful L1 material |
Unspent tokens are not compensated when an account is banned for breaching this Policy. Refunds of money are governed by the Refund Policy.
7. Appeals
The User may appeal any action by writing to support@iishki.com or Telegram @maksim_and_ai_support. An appeal is reviewed within five business days; if the action was applied in error, it is reversed and access is restored. An appeal does not suspend actions taken for L1 violations.