Skip to main content
Inpray

AI Use and Safety Policy

Effective date: 15 February 2026

This policy explains how AI features in Inpray are designed and how they should be used responsibly.

1. Purpose of AI Features

AI features are intended to support prayer, spiritual reflection, catechetical learning, and examination of conscience preparation in a respectful and non-judgmental way.

2. What AI Is Not

AI output is not sacramental confession, absolution, pastoral authority, therapy, medical care, legal advice, or emergency support. Users must consult qualified human professionals when needed.

3. Output Reliability

AI output may contain mistakes, omissions, or uncertain statements. Users should verify important points independently and should not rely exclusively on AI for high-impact decisions.

4. Safety Boundaries

The system is designed to avoid harmful, hateful, violent, sexual, exploitative, or illegal content. Abuse attempts, prompt injection, and evasion behavior may be blocked or logged for security review.

5. Crisis and Emergency Situations

If a user appears to be in immediate danger or in a mental health crisis, they must contact local emergency services or qualified crisis hotlines. The app is not a crisis-response service.

6. Human Escalation

For sacramental matters, users should speak with a priest. For health matters, users should contact licensed clinicians. For legal matters, users should seek qualified legal professionals.

7. Feedback and Continuous Improvement

User feedback (e.g., thumbs up/down) may be used to improve quality and safety over time. Feedback does not retroactively rewrite past answers for the current session.

8. Voice and Audio Behavior

AI voice playback is generated from text output. Audio controls (play, pause, stop, replay) are provided for usability. Voice output should be considered informational with the same limits as text.

9. Model and Provider Transparency

Current AI provider(s): OpenAI API. Models and configurations may change over time for safety, quality, and cost control.

10. Policy Updates

This policy may be updated as the product evolves. Material updates will be published in-app.