Google signs the non-mandatory code of conduct regarding European Union labeling obligations.
Google has signed the European code of conduct for transparency of AI-generated content. With this voluntary agreement, the company is preparing for the transparency obligations of the AI Act, which will come into force in the European Union over the coming years.
The code of conduct provides AI companies with a standardized framework to demonstrate that images, videos, audio, and text have been generated or edited by AI. Participation is not mandatory, but it helps companies prove they comply with European regulations.
SynthID and C2PA
Google is focusing on SynthID, a tool that adds invisible digital markings to AI-generated content. This metadata should clarify whether an image, video, audio file, or text was created or modified with AI. Additionally, Google supports the C2PA standard, an initiative that records the origin and editing history of digital content. The company is collaborating with Apple, Nvidia, OpenAI, ElevenLabs, and Kakao to normalize watermarking technology.
Companies that do not sign the code of conduct can still comply with the AI Act. They will have to demonstrate themselves that their systems for labeling AI content meet European requirements. By following the code of conduct, companies gain a recognized framework that can be used throughout the European Union as evidence of compliance with transparency obligations.
Google warns of label fatigue
Google also expresses reservations about the European approach. According to the tech giant, there is a risk that users will be overwhelmed with various AI labels and legal notifications when multiple platforms each use their own systems. According to the company, this could lead to more confusion instead of more transparency.
Google therefore emphasizes that transparency measures are only effective when users understand the labels and can trust them.