← Vetro Attest · Free Art. 12/14 checker · Free EU AI Act tools

EU AI Act Article 13: transparency for AI agents

Article 13 requires high-risk AI systems to be designed so that their operation is sufficiently transparent for deployers to interpret the output and use it appropriately. It also requires instructions for use in a concise, complete, correct and clear form.

What the instructions must contain

Among other things:

• The identity and contact details of the provider.
• The characteristics, capabilities and limitations of the system, including its intended purpose and the level of accuracy, robustness and cybersecurity it was tested for (see Article 15).
• Known or foreseeable circumstances that may lead to risks.
• The human oversight measures, including the technical measures that help deployers interpret output (see Article 14).
• The computational and hardware resources needed and the expected lifetime.
• Where relevant, a description of the mechanisms that allow deployers to properly collect, store and interpret the logs under Article 12.

That last point is easy to overlook. If you sell an agent to a deployer, the deployer has to keep its logs for at least six months (see Article 26). Your instructions must tell them how.

What "interpret the logs" means in practice

A deployer's compliance team will not read raw JSON. They need to answer:

• What did the agent do in this run, step by step?
• Was any action blocked or escalated to a human, and by which rule?
• Has anything in this log been changed since it was written?

A log format with a published schema, a public verification key and a verifier that runs outside your infrastructure answers all three. Include in your instructions for use the schema version, where the public key is published, and how to run verification.

A checklist for your instructions for use

• [ ] Log format and schema version named
• [ ] Where logs are written and how the deployer exports them
• [ ] How to verify integrity without contacting you
• [ ] Which events are recorded (model calls, tool calls, policy decisions, approvals)
• [ ] What the log does not show (for example, content you chose not to record)

Check your own logs

Paste a Vetro ledger or an OpenTelemetry GenAI trace export into the free checker: it verifies integrity and shows which record-keeping obligations your evidence covers, in your browser. When you need something to hand to an auditor or a customer, the countersigned EU AI Act report costs 29 USD, one-off: an ed25519 counter-signature over your evidence plus a mapping to the articles, including an explicit list of what the log cannot show.

See a sample ledger verified in one click, check your own logs free or get the countersigned EU AI Act report.