A new 5W AI Visibility Index audit found that when buyers ask the major AI engines who must disclose AI-generated advertising under the EU AI Act, the answer comes back from law firms and compliance-software vendors — and almost never from the communications industry.
The audit ran five buyer prompts — from “Does the EU AI Act require disclosing AI-generated ads?” to “Does the EU AI Act apply to US agencies?” — against the retrieval corpus that ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews draw from. Law firms including Sidley, Lewis Silkin, and Gleiss Lutz dominated. AI-governance vendors filled the rest. Not one global PR or communications firm surfaced on the compliance or scope prompts.
Two Clocks, One Deadline Season
The stakes are concrete. Article 50 of the EU AI Act splits across two dates. The core disclosure duties — deployers, including advertisers and brands, labeling deep fakes and AI-generated content — take effect August 2, 2026. The separate machine-readable watermarking requirement (Article 50(2)) was deferred by the Digital Omnibus to December 2, 2026, a four-month grace period for systems already on the market. Everything-PR broke down that watermarking timeline in The EU AI Act’s December 2 Deadline: What Marketers Must Know About AI Content Watermarking.



