AI Ethics & Governance Policy

1. Purpose & Scope

This policy establishes Mediar AI's commitment to responsible AI development and deployment. It applies to all AI systems developed, deployed, or used by Mediar AI, including:

• • Workflow automation agents
• • LLM-based assistants and orchestrators
• • UI automation systems
• • Any AI/ML models integrated into our products

2. Leadership Commitment

Mediar AI's leadership commits to:

• • Allocating resources for AI safety and compliance
• • Regularly reviewing AI system impacts and risks
• • Ensuring alignment between AI development and business objectives
• • Promoting continuous improvement in AI governance
• • Responding promptly to identified AI-related risks or incidents

3. Core AI Principles

3.1 Transparency

• • Users are informed when interacting with AI systems
• • AI decision-making processes are documented and explainable where feasible
• • Limitations of AI systems are clearly communicated

3.2 Human Oversight

• • Critical decisions maintain human-in-the-loop controls
• • Users can override or stop AI actions at any time
• • Automated workflows include cancellation mechanisms

3.3 Privacy & Data Protection

• • AI systems process only necessary data
• • Personal data is handled per GDPR requirements
• • User data is not used for model training without consent

3.4 Safety & Security

• • AI systems undergo security review before deployment
• • Prompt injection and adversarial attacks are mitigated
• • Credentials and sensitive data are never logged by AI systems

3.5 Fairness & Non-discrimination

• • AI systems are tested for bias before deployment
• • Discriminatory outputs are prohibited
• • Feedback mechanisms allow users to report issues

4. AI Risk Assessment

5. Governance Structure

As an early-stage startup, Mediar AI maintains founder-led governance with clear accountability:

As the company grows, an independent advisor with AI expertise will be appointed to provide external oversight.

6. AI System Inventory

7. Incident Response

AI-related incidents are handled through a structured process:

1. Detection — Monitoring alerts or user reports
2. Assessment — Evaluate impact and severity
3. Containment — Disable affected AI features if needed
4. Resolution — Fix root cause
5. Review — Document lessons learned

Report AI incidents to: matt@mediar.ai

8. Review & Updates

This policy is reviewed:

• • Annually (minimum)
• • After significant AI system changes
• • Following AI-related incidents
• • When regulations change (e.g., EU AI Act)