GOAL reports an outbreak-response AI layer lost intended functionality after data-protection changes.
Proposed mechanism: Design changes altered what the service could deliver.
Limit: One programme account; privacy rationale and patient effects not independently assessed.