AI agents | EU AI Act
Under the EU AI Act, logging means a high-risk AI system automatically records events over its lifetime so its operation can be traced and checked, and providers and deployers keep those logs.
High-risk AI systems must technically allow events to be recorded automatically over the lifetime of the system. The logs must make it possible to identify situations where the system may present a risk or undergo a substantial modification, to support post-market monitoring, and to monitor its operation.
Article 19 requires providers to keep the automatically generated logs under their control for a period suited to the system's intended purpose, of at least six months, unless other Union or national law says otherwise. Article 26(6) sets the same minimum for deployers, for the logs under their control.
Both, for the logs under their control. The provider builds logging into the system and keeps its logs. The deployer, the organization using the system, keeps the logs the system generates in its own operation, and monitors the system as the provider's instructions describe.
The logging duties apply to high-risk AI systems. An agent is high-risk when it is used in an area the Act lists, such as hiring, credit scoring, access to essential services, education or law enforcement, or when it is a safety component of a regulated product. Many agents are not high-risk, and other duties, such as the transparency rules in Article 50, can still apply. Take legal advice for your case.
They are complete, time-stamped, tied to a specific system and version, and shown not to have changed since they were written. The Act does not prescribe a technology, so the burden is on you to show the logs can be relied on.
TIP gives each logged agent action a signed, time-stamped record linked to the step before it, kept on a network of nodes, so a regulator can check that the logs are complete and unchanged. It is evidence that supports compliance, not a certificate of it, and your raw data stays on your systems.
Agent observability on TIP records each agent action as a signed, tamper-evident entry that anyone you choose can check, while your data stays on your systems. It is a commercial service from The AI Lab.
Related questions
