Build a board-ready risk heatmap in minutes — free, no account required

Open tool

Real-World Testing

High-Risk AI Real-World Testing Governance

What this control does

Plan, approve, register, govern and close real-world testing of high-risk AI under Articles 60/60a.

How to implement

Before real-world testing outside a sandbox, identify the applicable Article 60 or Article 60a route and responsible authority. Document the testing plan, necessary approval and registration, establishment arrangements, duration, participant protections, oversight and responsibilities. Obtain the required consent or assess any specific permitted alternative. Monitor the test, address incidents promptly and record suspension, termination and outcomes. For Annex I Section B, verify that the relevant Member State framework permits the testing and apply its incorporated safeguards; Article 60a is not a general EU-wide permission.

Suggested timing and triggers

Before testing; throughout the approved test; on incidents or material changes; at suspension, termination and completion.

Evidence examples

Testing-route decision and approved plan Approval, registration and establishment evidence Participant, oversight and responsibility arrangements Test duration, monitoring and incident records Suspension, termination and outcome notifications

How to check this control

Select a test and verify the route, approvals and actual operating conditions. Check that participant protection, oversight and incident escalation were active, and reconcile closure notifications with the test outcome.

Recognized by G2 as a Leader

Trusted by customers and rated highly across all categories

High-Risk AI Real-World Testing Governance | EU AI Act Suggested Control | Tracker Networks