AI usage declaration
You declare where AI is used, for which operational purpose, what outputs it generates, and its role in the process (assisted, automated, recommendation, support, etc.).
Transparent process
A simple, repeatable and defensible process. Designed to provide clarity, speed and verifiability without turning it into a heavy audit.
When required, the certificate can be reinforced through eIDAS-aligned timestamping issued by accredited external trust service providers.
You declare where AI is used, for which operational purpose, what outputs it generates, and its role in the process (assisted, automated, recommendation, support, etc.).
We define in writing what the certificate covers and what is explicitly excluded. This avoids ambiguity and misinterpretation.
We request sufficient minimum evidence: internal documentation, links, excerpts, configurations, examples or logs. Evidence requirements are proportional to risk and context.
A human accountable owner (role or position) is identified, assuming final responsibility for the certified AI usage.
We issue the certificate with a public verification ID and an integrity hash that allows anyone to verify the document has not been altered since issuance.
Timing depends on the case and evidence availability. The objective is to be fast without losing defensibility.
We will reply with the proposed scope and required minimum evidence.