Responsible AI Statement

At Haven, we use artificial intelligence to support people in building healthier, more conscious relationships. Because relationships are deeply personal, we believe AI must be designed and used with exceptional care.

This statement explains how we approach AI responsibly.

1. AI as Support — Not Authority

AI within Haven is designed to:

• support reflection and self-awareness
• offer perspectives, not instructions
• encourage emotional regulation and curiosity

Haven's AI does not replace human judgment, therapy, or professional care. Users remain in control of their decisions at all times.

2. Transparency & Clarity

We are transparent about the role of AI in Haven.

• Users are informed when they are interacting with AI
• We do not misrepresent AI as a human therapist
• We acknowledge that AI-generated responses may be imperfect

Our goal is trust through honesty — not illusion.

3. Emotional Safety by Design

Haven's AI is intentionally designed to:

• avoid blame, judgment, or shaming
• de-escalate emotional intensity
• promote empathy and perspective-taking

When sensitive topics arise (e.g. conflict, distress, relationship breakdown), the AI prioritises emotional safety and grounding.

4. Boundaries Around High-Risk Situations

Haven's AI:

• does not provide crisis intervention
• does not offer medical or mental health diagnosis
• does not encourage harmful or abusive behaviour

When signs of acute distress, abuse, or danger appear, Haven encourages seeking support from qualified professionals or trusted people.

5. Privacy & Data Stewardship

We use personal data responsibly.

• Relationship data is treated as highly sensitive
• AI interactions are protected by privacy safeguards
• We do not sell personal data
• We aim to minimise data use wherever possible

Where data is used to improve the system, it is anonymised and aggregated when feasible.

6. Avoiding Manipulation & Dependency

Haven's AI is not designed to:

• manipulate emotions
• nudge users toward specific decisions
• foster emotional dependency on the app

Our goal is to help users build skills and habits that strengthen real-world relationships beyond the platform.

7. Bias Awareness & Inclusivity

We strive to reduce bias in AI outputs by:

• using diverse perspectives in design
• reviewing language for assumptions and stereotypes
• continuously monitoring for unintended harm

We recognise that no system is perfect, and we commit to learning and improving.

8. Human Oversight & Continuous Improvement

Haven maintains human oversight of AI systems.

• Feedback loops allow users to report issues
• Outputs are reviewed and refined over time
• Safety and ethics guide product decisions

We see responsible AI as an ongoing practice, not a one-time decision.

9. Accountability & Feedback

We welcome questions, concerns, and feedback about how AI is used in Haven.

📧 hello@askhaven.app