AI Ethics Policy

Core Principles

Transparency: Dataset provenance, feature versions, model versions, evaluation reports, and limitations must be documented.

Fairness: Evaluation must include demographic subgroup analysis before any clinical-performance claim.

Human Oversight: Model output is decision support and never an autonomous medical decision.

Privacy and Purpose Limitation: Research use requires explicit consent, access control, encryption, and auditable retention rules.

Model Governance Status

• Production model: None approved
• Clinical validation: Pending independent evaluation
• Performance claims: Not established
• Promotion: Independent evaluation plus manual approval required

Reference Frameworks

EU AI Act, WHO AI Ethics Guidelines, OECD AI Principles

Contact

ethics@neuralcipher.ai