AI transparency controls: disclosure statements and image watermarking
Article 50 of the EU AI Act (Regulation (EU) 2024/1689) applies from 2 August 2026: people must be told when they are interacting with an AI system, and AI-generated media must be marked as such. Both are now enforced in path, so no application team has to implement them separately.
A policy-defined statement is injected into the conversation so the person on the other side knows they are talking to an AI assistant. The wording is set per policy, per language and per jurisdiction, and the exact text plus the policy version is written into the request record.
Images returned through the control plane carry a machine-readable provenance marker and optional visible mark, applied centrally rather than per provider. Coverage is the same whichever model produced the image, and the marking decision appears in the audit trail alongside redaction and routing.