ACCEPTABLE USE POLICY
Acceptable Use Policy
Applies to AI image generation, editing, uploading, storage, and sharing on Maixiang. Last updated September 8, 2026.
To report a violation, email zhexindx@gmail.com. We record and prioritize reports, take appropriate action, and share an outcome where appropriate.
On this page
1. Content standards and prohibited categories
You must not use the service to generate, edit, upload, store, or share content in these six categories:
Sexually explicit content
Pornography, sexually explicit material, or content intended for sexual exploitation.
Violence and gore
Content that incites, glorifies, or graphically depicts severe violence, injury, or abuse.
Hate speech
Content that demeans, dehumanizes, threatens, or incites discrimination against groups based on protected characteristics.
Child safety and CSAM
Any sexual exploitation or sexualization of minors, or other content that endangers children. We have zero tolerance for such content.
Deceptive deepfakes and impersonation
Misleading deepfakes, impersonation, or reputation-damaging depictions of real people created without their consent.
Copyright and trademark infringement
Unauthorized copying, adaptation, or misuse of material protected by copyright, trademark, or other intellectual property rights.
We also prohibit terrorist or extremist propaganda, instructions for making weapons of mass destruction, fraud, illegal activity, malware, deceptive impersonation, privacy violations, and jailbreak prompts designed to bypass safety checks. We have zero tolerance for child safety violations, terrorism, or non-consensual sexual deepfakes of real people.
2. AI models and generated content
We use third-party AI providers to generate and edit images. Providers and models may vary by tool and configuration. Content is generated automatically by these models. It is not human-authored by Maixiang or the model provider, nor does it carry their factual endorsement or commercial guarantee.
3. Enforcement
Our response depends on the severity of the content, account history, and whether the behavior is deliberate or repeated:
- Minor or first violations: a warning and a request to remove or change the content.
- Confirmed violations: removal of or restricted access to the generated content, with limits on related generation features.
- Repeated violations, high-risk behavior, or deliberate abuse: content removal, temporary account suspension, request limits, or removal of unused promotional credits or allowances.
- Serious illegal activity, child safety violations, terrorism, or non-consensual deepfakes: immediate blocking, permanent bans, and reports to relevant authorities where required by law or necessary.
If you believe an action was taken in error, you may appeal through the reporting email address. Someone who was not involved in the original decision will review the appeal.
4. Content moderation
We combine model-provider safety policies, platform rules, and human review. Submitted prompts and generated results are checked against provider safety policies and our own rules. Blocked, flagged, or reported content is referred for human review. We regularly sample resolved reports and improve our rules. Confirmed violations may lead to warnings, content removal, generation limits, account suspension, or permanent termination of access. We cooperate with relevant authorities on suspected illegal content as required by law.
4.1 Automated checks and human review
- Input checks identify clear policy violations, illegal intent, and attempts to bypass safety measures in prompts.
- Output checks assess generated results using provider safety policies and platform rules. Results that fail these checks are not delivered to users.
- Human review covers reports, system flags, and high-risk content. Administrators review this material and keep records of their decisions.
- Abuse monitoring looks for bulk generation, repeated violations, unusual request patterns, and account misuse.
| Severity | Examples | Response |
|---|---|---|
| Critical (L1) | Child safety, terrorism, or mass violence | Block immediately, verify manually, and ban accounts or report to authorities where necessary and required by law |
| High risk (L2) | Serious violations or clear infringement | Prioritize review, remove content, and suspend relevant accounts |
| Standard (L3) | Confirmed standard violations | Warn the user, remove content, and restrict features if needed |
| Low risk (L4) | Minor violations or reports needing more information | Review manually, request changes, and record the outcome |
4.2 Response timelines
| Severity | Initial response | Target resolution |
|---|---|---|
| L1 Critical (child safety, terrorism, etc.) | Within 2 hours | Within 24 hours |
| L2 High risk | Within 24 hours | Within 3 business days |
| L3 Standard violations | Within 3 business days | Within 7 business days |
| L4 Low risk or incomplete information | Within 5 business days | Within 15 business days |
Our process is to receive and record a report, assign an initial priority, run automated checks and human review, decide on action, carry it out, and share the outcome where appropriate. If more time is needed, we will explain why.
5. Reporting and appeals
If generated content violates this policy or your rights, or involves serious concerns such as child safety, terrorism, or deepfakes, email zhexindx@gmail.com. This is our staffed contact address for content safety concerns, reports, and appeals.
Use “Content report” as the subject line and include the generated image link or job ID, violation category, description, and relevant screenshots where possible. Reporter identities and contact details are kept confidential and used only as needed to handle the report, maintain safety, or meet legal obligations.
We record and acknowledge reports, assign a priority, and carry out automated checks and human review. We then decide whether to retain, restrict, or remove content, ban an account, or report a matter as required by law, and update the reporter where appropriate. Malicious, duplicate, or clearly false reports are also recorded and addressed. You may appeal a decision using the same email address. Appeals are reviewed by someone who was not involved in the original decision.
6. Related policies
This policy applies together with our Terms of Service and Privacy Policy. We update this policy to reflect model capabilities, legal requirements, and operational practices.