Google has signed the EU’s voluntary Code of Practice on Transparency of AI-Generated Content, aligning itself with a framework designed to help companies meet the AI Act’s rules for marking and labeling synthetic material.
The move comes ahead of Article 50 transparency obligations taking effect from 2 August 2026. Under those rules, providers must make AI-generated or manipulated content detectable in a machine-readable format. Deployers, meanwhile, must label deepfakes and certain AI-generated publications concerning matters of public interest.
Although joining the Code is voluntary, the transparency requirements in the AI Act itself are legally binding. Companies that do not sign the Code must show through other measures that their systems still comply with the marking and labeling rules.
Google said it will build its approach on SynthID, its technology for embedding digital watermarks in AI-generated text, images, audio, and video. The company is also backing the C2PA content-provenance standard and is working with Apple, ElevenLabs, Kakao, Nvidia, and OpenAI on interoperable watermarking tools using SynthID.
At the same time, Google warned that overlapping labels and legal disclosures could confuse users rather than make information clearer. It also said that because technical solutions are still evolving, extra regulatory complexity could work against the EU’s competitiveness and simplification goals.
Digital watermarking and responsible AI frameworks are becoming key parts of broader efforts to strengthen trust and accountability in online content.
Google’s participation gives added weight to the Code as an EU-wide implementation framework before the Article 50 rules come into force. Its practical effect, however, will depend on whether watermarking, provenance records, and visible labels stay detectable and understandable as AI-generated content moves across different tools and platforms.
