Google Signs EU Code on AI-Generated Content Transparency

Google has officially signed the EU's "Code of Conduct on AI-generated content transparency," developed within the framework of the EU's Artificial Intelligence Act. This initiative aims to foster the responsible use of artificial intelligence in the European market.

The new code of conduct outlines key requirements. These include machine-readable identifiers and watermarks to automatically detect content generated or modified by AI. It also mandates clear notification when AI systems interact directly with humans, ensuring users are aware they are communicating with a machine. For "deepfakes," content publishers must explicitly declare that the material is AI-generated or modified. AI-generated text intended for public interest information, such as news reports or policy interpretations, must also be clearly labeled.

Google states that this commitment builds upon its previous adherence to the "Partnership on AI Code of Conduct" signed in 2025. The company has also been actively promoting the C2PA standard for content provenance and developing its own AI transparency tools, such as SynthID.

The company intends to continue promoting its SynthID digital watermarking technology to help users identify how content is generated and edited. Google is also collaborating with third-party AI entities, including Apple, ElevenLabs, Kakao, Nvidia, and OpenAI, to encourage industry-wide adoption of interoperable watermarking tools based on SynthID.

While advocating for transparency, Google has also expressed concerns regarding increased regulatory complexity. The company suggests that adding more intricate regulations, especially while technological solutions are still evolving, could hinder Europe's pursuit of competitiveness and streamlined processes. Google warns that an internet flooded with overlapping AI labels and legal declarations could confuse the public and impede access to clear, necessary information.