Policy AI

AI Use Policy ​

Last updated: January 20, 2026

Purpose ​

This policy explains principles, scope, roles, and user rights related to AI systems used by our products and services.

Scope ​

Applies to all AI models, datasets, integrations, and features developed, procured, or maintained by the organization.

Principles ​

• Safety and reliability: We evaluate models for accuracy, robustness, and failure modes before deployment.
• Fairness and non-discrimination: We test for biased outcomes and take steps to mitigate disparate impacts.
• Transparency: We provide clear user-facing explanations of AI capabilities and limitations.
• Accountability: Owners and reviewers are assigned for each AI feature and decision process.
• Privacy: Data used for model training and inference is handled according to our privacy controls.

Governance & Roles ​

• AI Risk Committee: Reviews high-risk models and approves exceptions.
• Model Owner: Responsible for lifecycle, testing, monitoring, and documentation.
• Privacy & Legal: Assess data usage, regulatory compliance, and contractual obligations.

Development & Deployment ​

1. Requirement scoping: Define desired behaviors and success metrics.
2. Data preparation: Document sources, consent, and quality checks.
3. Testing: Unit, integration, bias, adversarial, and safety tests.
4. Staged rollout: Canary or phased deployment with monitoring.
5. Post-deploy monitoring: Track performance, drift, and incidents.

User Rights & Controls ​

• Explainability: Users receive plain-language information about automated decisions affecting them.
• Opt-out: Where feasible, users can opt out of automated profiling or personalized decisions.
• Correction: Users can request correction of incorrect data used by models.

Incident Response ​

We maintain an incident playbook for model failures, harm reports, or data breaches, including rollback and communication plans.

Documentation ​

Each model must have a Model Card and Data Sheet with purpose, training data summary, evaluation metrics, limitations, and contact points.

Contact ​

For questions or to report issues, contact the AI Governance team at support@avnlearn.com.