EU AI Act
1. Our position
Brayns builds AI for regulated compliance work, so we treat the EU AI Act as a design requirement rather than a future problem. This page describes how the platform is built to meet the Act's requirements as they phase in.
2. Risk management
The platform is developed under a documented risk-management process covering model behaviour, data handling and operational failure modes, reviewed as the system and the regulation evolve.
3. Transparency and documentation
Model versions are named, dated and explained. Decisions carry reasoning that can be examined, and technical documentation is maintained to support customer and regulator scrutiny.
4. Human oversight
The platform is built for oversight by design: uncertain cases escalate to your people, judgement stays with the humans accountable for it, and every action is on the record.
5. Data governance
Training and evaluation data for the model are governed by documented processes. Customer data is not used to train models, and deployments run inside the customer's own perimeter.
6. Standards alignment
Our AI management system is certified under ISO/IEC 42001:2023, the international standard most closely aligned with the Act's management-system expectations, alongside our information-security and privacy certifications.
7. Contact
Questions about our AI Act work can be sent to contact@brayns.ai.