Acceptable Use Policy

Prohibited uses

You may use I'm Sorry I Lied AI only for lawful purposes and content you are authorized to process. You must not use the service to create, upload, request, share or facilitate:

• Content or conduct that violates applicable law or assists unlawful activity.
• Use of another person's image, voice, identity or likeness without the authorization required for the requested processing.
• Material that infringes copyright, trademark, privacy, publicity, likeness or other intellectual property or personal rights.
• Non-consensual intimate imagery, unlawful sexual content, or sexual content involving a person who has not consented.
• Any sexualization, exploitation or endangerment of a minor, including child sexual abuse material or simulated depictions.
• Content depicting graphic violence, extreme gore, mutilation, torture, dismemberment, severe bodily injury, or other excessively bloody or disturbing imagery. Content that glorifies, encourages, or promotes real-world violence is strictly prohibited.
• Hateful, discriminatory, or degrading content targeting individuals or groups based on race, ethnicity, nationality, religion, disability, sex, gender, gender identity, sexual orientation, or other protected characteristics. This includes hateful slurs, dehumanizing portrayals, harassment, incitement to hatred, and advocacy of discrimination or violence.
• Fraud, scams, phishing, deceptive impersonation, harassment, defamation, targeted abuse or misleading content intended to cause harm.
• Attempts to bypass safeguards, misuse access credentials, overload the service, evade restrictions or access the service without authorization.

Content Moderation and Reporting

I'm Sorry I Lied AI operates automated content moderation checks on both uploaded images and AI-generated results to enforce this policy. These checks use third-party safety detection services and run before content is processed or delivered.

Automated Content Moderation

All uploaded images are scanned using Replicate's NSFW image detection service before being submitted for AI generation. Generation requests are blocked when the upload does not pass this check.

Server-side prompts are scanned using Waffo's content safety API before generation begins. Requests that fail the prompt safety check are blocked or held for review.

The AI generation provider (SeeAPI) applies its own content safety filtering during image creation using the Qwen-Image-Edit model's built-in safety mechanisms.

All generated images are scanned using Replicate's NSFW image detection service before being stored and delivered to users. Results that do not pass this output check are blocked from delivery, and any reserved credits are automatically refunded.

When any moderation service is temporarily unavailable, the request is blocked rather than bypassed. Failed moderation does not result in automatic approval.

Human Review and User Reports

Users may report content that may violate this policy by emailing [email protected]. Reports should include the generation task ID or image link (if available), the reason for the report, and any additional context that may assist review.

Reportable concerns include: AI-generated images that appear to violate this policy; unauthorized use of a person's likeness; intellectual property infringement; content involving minors; or attempts to bypass safety checks.

All reports are reviewed by the site operator. The reviewer examines the reported content, associated generation records, and moderation logs where available. Reporter identity is not disclosed in the review process.

Content Severity and Response Times

Reports are categorized by severity and addressed within the following targets:

These targets represent our good-faith commitment. Actual resolution time may vary based on report complexity, volume, and the need for additional verification. Response times begin when a report is received at [email protected] during normal operating hours.

• L1 — Critical (child safety, non-consensual intimate imagery): Initial response within 2 hours, target resolution within 24 hours.
• L2 — High (copyright infringement, identity misuse, violent content): Initial response within 24 hours, target resolution within 3 business days.
• L3 — Medium (hate speech, discrimination, other policy violations): Initial response within 3 business days, target resolution within 7 business days.
• L4 — Low (general inquiries, borderline content): Initial response within 5 business days, target resolution within 15 business days.

Enforcement Actions

When a policy violation is confirmed, appropriate action is taken based on severity:

• Blocking the generation request or removing the delivered content.
• Refunding credits for blocked or removed content.
• Restricting the user's access to generation features.
• Suspending or terminating the user's account for repeated or severe violations.
• Preserving relevant records for legal compliance or dispute resolution.
• Reporting conduct to authorities where required by law.

Enforcement

We may reject generation requests, block or remove prohibited content, restrict access to our services, or suspend or terminate accounts when violations of this Policy are identified. We may investigate reported violations and take appropriate action based on the severity and circumstances of each case.

Reporting violations

If you encounter content or activity that may violate this Policy, please contact us using the support email listed on our website. Include relevant details so we can review the report and take appropriate action.

Responsibility and enforcement boundary

Automated moderation checks reduce but do not eliminate risk. No automated system detects every prohibited item with complete accuracy. You remain responsible for your submissions and how you use generated results.

The service blocks requests before generation when input or prompt checks fail, and blocks delivery when output checks fail. Moderation logs are retained for operational and compliance purposes. Reports sent to [email protected] are reviewed as reasonably practicable.