AI Usage in Kahf Products — Responsible & Halal AI

Our AI Philosophy

At Kahf, every technology decision is filtered through two lenses: Islamic ethics and user benefit. AI is a powerful tool, but one that requires careful stewardship. We apply it only where it creates genuine value, never in ways that compromise your privacy, dignity, or faith.

Privacy First

On-device AI processing is our default. Your data stays on your device whenever possible.

Transparency

We tell you when AI is being used and what it does. No hidden profiling.

Human Oversight

AI assists — it does not decide. Humans remain in control of every significant outcome.

Purposeful Use

We only adopt AI where it solves a real problem for our users, never for hype.

AI in Our Products

Specific, honest detail on how AI is applied in each Kahf product.

Kahf Browser

On-Device AI

Kahf Browser uses on-device AI to detect and blur indecent imagery as you browse — in real time, without sending any images to our servers. The model runs entirely on your device, ensuring that your browsing remains private.

AI is used for

On-device image classification to detect indecent content
Automatic blurring before images are rendered on screen
Continuous learning updates shipped via app updates, not user data

AI is NOT used for

Sending browsing history to Kahf servers
Building advertising profiles
Cross-device tracking

Kahf Guard (SafeGaze)

Filter AI

Our browser extension uses AI-powered domain classification to identify and block harmful, anti-Islamic, and privacy-invasive content. The blocklist is maintained by our team and refined continuously using machine-learning-assisted classification.

AI is used for

Domain and URL classification into harm categories
Detection of tracking and fingerprinting scripts
Adaptive filter updates based on emerging threats

AI is NOT used for

Reading the content of pages you visit
Collecting personally identifiable data
Profiling individual users

KahfTube (Kahf TV)

Content AI

KahfTube leverages AI to classify YouTube channels and videos according to Islamic content guidelines. The classification engine runs on our servers against public metadata — never against your watch history.

AI is used for

Channel classification by content type and Islamic permissibility
Video-level metadata analysis (title, description, tags)
Automatic detection of newly uploaded haram content

AI is NOT used for

Analyzing your personal watch history
Storing or selling viewing patterns
Accessing private account information

Kahf Kids

Safety AI

Kahf Kids employs multi-layer AI content filtering to ensure the platform remains safe for children. Every piece of user-generated or third-party content passes through automated safety checks before being displayed.

AI is used for

Age-appropriate content classification
Real-time moderation of user-generated content
Detection of grooming patterns and inappropriate interactions

AI is NOT used for

Collecting children's personal data beyond what is legally permitted
Behavioral advertising targeting children
Sharing children's data with third parties

Mahfil

Moderation AI

As a halal video-sharing platform, Mahfil uses AI-assisted moderation to review uploaded content for violations of Islamic community guidelines — significantly reducing the workload on human moderators.

AI is used for

Automated pre-screening of uploaded videos
Flagging content for human moderator review
Detecting known categories of impermissible material

AI is NOT used for

Automatically removing content without human review for edge cases
Monitoring private messages
Influencing content ranking based on commercial relationships

Our AI Safeguards

Every AI system at Kahf is governed by a set of binding safeguards to protect your rights and our shared Islamic values.

No Training on User Data

We do not use your personal data, browsing history, or content to train or fine-tune our AI models without explicit opt-in consent.

Regular Bias Audits

We periodically audit our classification models for cultural and religious bias, particularly to ensure Islamic content is treated fairly.

Right to Contest

If an AI decision affects you (e.g., content incorrectly blocked), you can report it and a human will review the decision.

Minimal Data Retention

Any data processed by cloud-based AI components is retained only as long as necessary and deleted per our data retention policy.

What We Will Never Use AI For

Predicting or profiling religious beliefs or political views
Autonomous financial decisions without explicit user action
Manipulative or deceptive user interfaces (dark patterns)
Surveillance of users' physical locations or activities
Selling AI-derived insights about users to advertisers
Replacing human judgment in matters of Islamic ruling (fatwa)

Frequently Asked Questions

Does Kahf use Large Language Models (LLMs)?

Some of our research and internal tooling uses LLMs for productivity. None of our user-facing products currently use LLMs to process your personal data. When that changes, we will update this page.

Who reviews the AI classifications?

Our in-house content team — composed of practicing Muslims with knowledge of Islamic jurisprudence — provides ongoing human oversight of AI classification decisions, particularly for edge cases.

Can I opt out of AI-powered features?

Most AI features are integral to the protection offered by our products. However, where technically feasible, we provide controls in the app settings to adjust or disable specific AI-driven behaviors.

How do you ensure AI does not introduce religious bias?

We maintain a diverse team that includes scholars and community members from different Islamic traditions. Our classification guidelines are reviewed by qualified scholars to ensure they respect the diversity of the Muslim Ummah.

Questions about our AI practices?

We welcome questions, feedback, and accountability from our community.

Contact Our AI Team