Astronia — Powered by AP Pure Care Private Limited

AI Transparency Policy

How Astronia uses AI, and its limits

• Effective date: 6 September 2026
• Version: 2.2
• Purpose: Controls for generative astrology, coaching and support features
• Applies to: Users, employees, vendors and Service Providers interacting with AI

1. Clear identification

Every automated conversational feature is clearly identified as an AI system before the User begins. The interface must not state or imply that an AI agent is a human astrologer. Synthetic voice, avatar, image or video is labelled where required by law and wherever a reasonable user could otherwise be misled.

2. Limitations

AI responses may be incomplete, incorrect, inconsistent or generated from patterns rather than verified facts. They are for entertainment, reflection and general information, not professional advice or guaranteed prediction. Users should verify important information and consult a qualified human professional where appropriate.

3. Prohibited AI outputs

• A guaranteed prediction of death, disease, pregnancy, fertility, marriage, divorce, litigation, visa, employment or investment return.
• Medical diagnosis, medication instruction, legal strategy, personalised securities recommendation or crisis counselling presented as professional advice.
• Any claim that a remedy, gemstone, yantra or ritual diagnoses, cures, mitigates, treats or prevents a disease or condition (the Astrology & Outcome Disclaimer, Cl. 3.2).
• Fear-based claims that the User is cursed, possessed or in imminent supernatural danger, especially when linked to payment.
• Requests for passwords, OTPs, card details, bank transfers, intimate images or off-platform payment.
• Discriminatory, sexually exploitative, violent, self-harm encouraging or child-endangering content.

4. Human review and escalation

Astronia may use automated safety filters and trained reviewers. High-risk prompts involving self-harm, violence, abuse, medical emergency, fraud or child safety trigger a safety response and, where appropriate, human escalation. AI escalation does not create an emergency-service relationship.

5. User input and model improvement

Users must not submit another person’s sensitive data without authority. Identifiable private Consultations are not used for general model training without separate consent. De-identified quality evaluation may be used where reasonable safeguards prevent re-identification.

6. Synthetic-content provenance and user declarations

Where Astronia enables creation, modification, upload, transmission or publication of synthetic audio, video or images, it deploys reasonable and appropriate technical measures, including automated tools or other suitable mechanisms, to prevent unlawful synthetic content. For permissible synthetic content, Astronia requires an appropriate User declaration, applies a clear and prominent visual label or prefixed audio disclosure, and embeds permanent metadata or another technically feasible provenance mechanism, including a unique identifier for the relevant Astronia computer resource where required. The Platform does not enable removal or suppression of a required label, metadata or identifier, and Users must not attempt to remove, conceal or defeat them.

7. Safety records and appeals

Astronia records material automated restrictions, gives a reasoned appeal route and preserves evidence required for lawful investigation. AI safety controls supplement, and do not replace, human grievance review.