Content Moderation Policy

1. Purpose & Scope

qibla.ai is an educational AI platform that provides factual, balanced information about Islam. This policy describes how we moderate content to ensure accuracy, safety, and respect for Islamic scholarly traditions. It applies to all AI-generated responses and user interactions on the platform.

2. What We Filter (Input)

User queries are screened before being processed. The following categories are redirected to safety responses:

• Violence & Extremism: Queries that seek to justify violence, recruit for extremist groups, or request instructions for harmful acts. These are redirected to content about Islam's emphasis on peace, mercy, and the sanctity of life.
• Sectarian Hatred: Queries designed to generate hatred between Islamic schools of thought or declare groups as non-Muslim. These are redirected to content about Islamic unity and scholarly diversity.
• Self-Harm: Queries expressing intent to self-harm are met with crisis resources (988 Suicide & Crisis Lifeline, Samaritans, Crisis Text Line) alongside compassionate Islamic guidance on patience and hope.
• Fatwa Requests: Explicit requests for binding religious rulings are redirected with an explanation that qibla.ai does not issue fatwas, and users are encouraged to consult qualified scholars.

Legitimate educational questions about sensitive topics (e.g., the history of conflict, comparative jurisprudence) are processed normally with additional scholarly context.

3. What We Filter (Output)

AI-generated responses are validated before being shown to users. The AI is instructed to never:

• Issue definitive religious rulings (fatwas) — responses present scholarly opinions, not binding edicts
• Present one school of thought as the only correct view on legitimately disputed matters
• Generate content that could be used to justify violence
• Reproduce copyrighted translations verbatim
• Make claims about individuals' faith status (takfir)
• Dismiss or mock any legitimate Islamic school of thought

The AI is required to always:

• Acknowledge when a question involves legitimate scholarly disagreement
• Present multiple scholarly opinions on disputed fiqh matters, labeling which school or scholar
• Distinguish between Quran (divine text), authenticated Hadith (prophetic tradition), and scholarly opinion (ijtihad)
• Include “والله أعلم” (And Allah knows best) on interpretive answers
• Cite specific verse/hadith numbers with precise references

4. Review Process

Flagged content is reviewed through our administrative moderation system, which uses AI-assisted toxicity detection alongside human review. Content is categorized by severity and type, with suggested actions ranging from dismissal (for legitimate educational queries) to restriction (for clearly harmful content).

5. Reporting Issues

Users can report inaccurate or problematic responses using the Report Inaccuracy button (flag icon) available on every AI response. You can also contact us directly:

• Email: content@qibla.ai
• In-app feedback buttons (thumbs up/down on every response)

6. Appeals

If your query was redirected and you believe it was a legitimate educational question, you can rephrase your question with more specific academic context. If you believe a moderation decision was made in error, contact us at content@qibla.ai and we will review within 5 business days.

7. Updates & Transparency

This policy is reviewed quarterly and updated as our moderation systems evolve. Significant changes will be communicated through the platform. We are committed to transparency about our methodology and limitations.