Ethics Guidelines

Responsible AI and Ethics at Kayalogy

We build medical AI with a safety-first mindset. Our approach is grounded in respect for people, clinical responsibility, data protection, and scientifically sound methods.

Intended Use and Human Oversight

• Non-clinical use: During development, Kayalogy is provided for research and evaluation purposes only.
• Human oversight: We design for appropriate human control and do not aim to replace clinical judgment.
• Transparency: We communicate intended use, limitations, and validation status clearly.

Core Ethical Principles

• Patient benefit: We prioritize patient well-being, safety, and clinical usefulness.
• Privacy and confidentiality: We minimize access to sensitive data and apply privacy-by-design.
• Fairness: We aim to reduce bias and assess performance across relevant populations where feasible.
• Accountability: We document decisions, evaluate risks, and improve controls as we scale.

Clinical and Research Integrity

• We seek alignment with Swiss law and relevant international ethical frameworks.
• Where research involves human data, we follow informed-consent principles and recognized guidance (e.g., the Declaration of Helsinki) where applicable.
• We aim to use robust validation methods, maintain auditable records, and avoid overstating performance.

Quality, Safety, and Risk Management

• Safety-by-design: We treat safety as a product requirement, not a feature.
• Risk management alignment: We aim to align with established practices (e.g., ISO 14971) as the platform matures.
• Quality mindset: We work toward quality-management discipline commonly used in MedTech (e.g., ISO 13485-aligned processes), appropriate to development stage.

Governance and Conduct

• Conflict of interest: We identify, disclose, and manage potential conflicts.
• Anti-corruption: We do not tolerate bribery, improper influence, or unethical business practices.
• Security mindset: We continuously improve safeguards and access controls.
• Reporting: We encourage early reporting of concerns and investigate issues promptly.

Model Lifecycle and Continuous Improvement

• We monitor performance and quality signals during development to identify issues, drift, and failure modes where feasible.
• We aim to implement documented review processes for updates, change control, and incident response.

Engagement and Feedback

• We value dialogue with clinicians, researchers, patients, and partners to improve safety and trust.
• Ethics and safety feedback: [email protected]

Important: Kayalogy is currently in development. Any previews or demonstrations are provided for evaluation purposes only and are not intended for clinical use.