Services · In preparation
AI Audit
Auditing high-risk artificial intelligence, when the regulation requires it.
The European AI Act establishes that high-risk artificial intelligence systems must undergo conformity assessments carried out by independent bodies. We are preparing for it.
CerteIDAS took part in the Spanish AI regulatory sandbox, the first controlled testing environment in Europe, where we worked with regulators and companies on the procedures for auditing AI systems. At ETSI we contribute to the working groups defining the technical standards for the cybersecurity audit of AI systems under the AI Act.
Scope
What we will assess
The scope will be set by the harmonised standards and the implementing acts of the AI Act, although we can already anticipate the following areas:
- High-risk AI systems (Annex III of the AI Act) subject to third-party conformity assessment;
- Risk management system, data governance, technical documentation and record-keeping;
- Transparency, human oversight, accuracy, robustness and cybersecurity of the system;
- The provider’s quality management system and post-market monitoring.
What sets us apart
We are where the standards are drafted, where the execution of the AI Act is designed.
We took part in the Spanish AI regulatory sandbox and in the ETSI groups defining the standards for the cybersecurity audit of AI systems, and we are members of the AI Act Advisory Forum. When the standards mandated by the European Commission are ready, so will we be.
Why it is not yet a service
We do not sell what cannot yet be certified.
We want to assess the conformity of AI systems with the AI Act. But the technical standards against which to run that assessment do not yet exist: not for anyone. That is why we are where they are being written.