AIGC Acceptable Use Policy

01Scope

This Policy applies to image generation, prompts, reference-image uploads, generation history, user-configured image APIs, and future generative-AI features offered through xiaoxiaodong.ai, www.xiaoxiaodong.ai, and vip.xiaoxiaodong.ai (collectively, “AIGC Features”).

It forms part of the Terms of Service and should be read with the Privacy Policy. More specific AIGC safety rules control over general terms for the same subject, without displacing mandatory law.

02User responsibilities

• You are responsible for prompts, reference images, instructions, outputs, and downstream use, and must have all rights, consents, and permissions needed to submit them.
• You may not deceive, jailbreak, or bypass our or a provider’s safety classifiers, restrictions, watermarks, content credentials, rate limits, or risk controls.
• Outputs can be inaccurate, biased, or include third-party material. Apply meaningful human review and obtain necessary permissions before publication, commercial use, or use in medical, legal, financial, employment, or other consequential decisions.
• Where law requires synthetic-media disclosure, apply a clear label and do not remove content credentials or generation labels that must be retained.

03Prohibited content and activity

You may not use AIGC Features to create, edit, upload, distribute, optimize, or facilitate:

• content that violates law, threatens national or public security, or promotes terrorism, violent extremism, or violent crime;
• child sexual abuse material, sexualization or exploitation of minors, sexual violence, non-consensual intimate imagery, abuse, or trafficking;
• instructions or encouragement for suicide, self-harm, eating disorders, or conduct creating an imminent risk of serious bodily harm;
• hate, dehumanization, unlawful discrimination, harassment, or violence directed at a protected class;
• fraud, phishing, fabricated evidence, false endorsements, extortion, market manipulation, malware, unauthorized intrusion, or materially facilitating unlawful conduct;
• deceptive impersonation or deepfakes of a real person without appropriate consent where likely to mislead, defame, harass, extort, or infringe privacy, publicity, voice, or likeness rights;
• infringement of copyright, trademarks, patents, privacy, publicity, reputation, personal-information rights, trade secrets, or other legal rights;
• unlawful surveillance, inference of sensitive personal data, indiscriminate biometric identification, or high-risk decisions made solely from protected or sensitive characteristics;
• conduct that interferes with the Service, abuses compute or provider quotas, or attempts to discover or defeat safeguards; or
• other content with substantially similar risk or a clear likelihood of serious harm to a person, society, or the Service.

04Reporting violations and emergencies

Any person may report suspected policy violations, rights infringement, or real-world safety risks through either channel below:

Please provide the exact URL, work/task/account identifier or location; the policy category and reasons; screenshots, timestamps, and verifiable evidence; the requested action; and contact details if follow-up is needed. A rights complaint should also include claimant identity, proof of rights, and an accuracy statement. Do not send passwords, payment passwords, full API keys, or unrelated sensitive information.

05Content-moderation process

We use risk-based layers where the relevant feature and available tools permit. We do not promise to identify every violation before generation, and moderation does not replace the user’s responsibility. Our process may include:

1. Request review: prompts, reference images, account behavior, and request patterns may be assessed using rules, automated safety models, or risk signals. A request may be blocked, rate-limited, or returned for revision.
2. Generation safeguards: we may configure or rely on a selected provider’s classifiers and output restrictions. A user-configured API remains subject to this Policy and the provider’s rules.
3. Post-generation review: user reports, complaints, safety logs, anomaly patterns, or legal requests may trigger review. High-risk, context-dependent, or uncertain cases may be escalated for human review.
4. Risk assessment and decision: we consider the nature of the content, real-world harm, distribution, intent, violation history, user capability, available evidence, and applicable law before deciding to allow, restrict, remove, or take account action.

Authorized reviewers may access relevant inputs, outputs, account data, and logs only as reasonably necessary for safety, report handling, troubleshooting, or legal compliance, subject to the Privacy Policy, role-based access, and retention controls.

06Enforcement measures

Depending on severity and risk, we may refuse or stop a generation, isolate or remove content, reduce frequency, restrict a feature, require verification, warn a user, suspend or terminate an account or membership, prevent re-registration, preserve evidence, notify an affected rights holder, or report to a competent authority.

We may first request correction for a minor, remediable issue. Child sexual exploitation, imminent danger, terrorism, serious fraud, systematic infringement, malicious safeguard evasion, or repeated violations may result in immediate restrictions before notice or full review.

07Notice, appeal, and review

Unless notice would obstruct an investigation, increase harm, expose safeguards, or violate a legal or authority restriction, we will try to tell an affected user the type of action, principal reason, and available appeal channel.

A user or reporter may appeal through the channels in Section 4 by identifying the account and original decision or report, explaining the objection, and supplying new evidence. An appropriately authorized reviewer will reconsider the available record and, where reasonably possible, will not rely only on the original automated result. We will respond within a reasonable period with a decision to uphold, modify, or reverse the action. An appeal does not automatically pause an emergency safety restriction.

08Privacy, transparency, and confidentiality

Personal information involved in moderation and reports is handled under the Privacy Policy. A reporter’s contact details are not ordinarily disclosed to the reported person unless disclosure is legally required, the reporter consents, or it is necessary for a legal proceeding. The reporting mechanism may not be abused to harass others, submit fabricated evidence, or suppress lawful content.

09Policy changes

We may update this Policy for changes in law, product capability, provider rules, or risk. The updated date will be posted here. A material change to user rights or obligations will be communicated through a prominent site notice, email, or another reasonable channel.

10Contact us