Parks must mark AI-generated promotional images before they go public. Which visual policy control should they apply?

Watermarking is a visual policy control that marks AI-generated media so provenance is visible at publication time. Parks releasing promotional images produced by generative models should apply watermarks before those assets go public, signaling that the creative originated from the city’s generation pipeline and supporting disclosure or brand policy. Watermarks complement—not replace—review workflows and content records.

Apply audit logs to generated promotional images before public release fails the visual policy control because logging creates backend records instead of adding a visible provenance mark, so audiences won't see AI origin at a glance. Apply content credentials to generated promotional images before public release fails the visual policy control because credentials embed provenance metadata rather than adding a visible mark, so viewers looking at the image itself won't see AI origin. Apply provenance tags to generated promotional images before public release fails the visual policy control because tagging attaches provenance data instead of marking provenance visibly and does not signal AI origin to the audience at publication time.

Exam caveat: watermarks aid disclosure and tracing; determined removal is possible, so do not rely on provenance tags alone, which attach provenance data instead of providing a visible mark of AI provenance. Some channels may require additional text disclosure besides a visual mark. Operational check: apply the approved watermark template to generated promotional images, verify visibility on typical display sizes, publish only watermarked finals, and do not substitute provenance tags for disclosure, but remember that provenance tags attach provenance data rather than marking AI provenance visibly.