In a major step toward improving transparency in artificial intelligence, the European Commission is intensifying efforts to create a Code of Practice that will mark and label AI-generated content. This initiative comes as the use of generative AI technologies, such as synthetic media and deepfakes, is on the rise, prompting essential discussions about accountability and consumer protection.
Industry and Civil Society Collaboration
The latest workshops convened by the European Commission's AI Office gathered a wide range of stakeholders, including technology companies, generative AI providers, civil society organizations, and academic experts. These discussions are key as they will shape the final draft of the transparency code, which is expected to be completed by early June. The code aims to support obligations under the EU's broader AI Act, which includes clear requirements for marking, labeling, and disclosing AI-generated content.
Key Focus Areas of the Working Groups
Two working groups were formed to address specific aspects of the transparency obligations. Working Group 1 focused on the marking and detection responsibilities of AI providers. Participants explored a revised multi-layered approach to compliance, raising concerns about feasibility and potential impacts on innovation. Civil society representatives, however, pushed for stronger protections to serve the public interest.
Meanwhile, Working Group 2 examined the disclosure requirements for those deploying generative AI systems, particularly regarding deepfakes and AI-generated texts. Key topics included the need for origin disclosure, user-facing labels, and the potential for a standardized EU label to unify these disclosures across member states.
Technical Challenges and Regulatory Balance
The workshops also looked at integrating machine-readable marks, provenance data, and watermarking systems throughout the AI value chain. Participants stressed the importance of coordination with existing EU regulations, such as the Digital Services Act, while recognizing the challenge of balancing transparency with legal clarity and innovation.
Developing the code presents its own set of challenges. Stakeholders are handling the complexity of transforming theoretical transparency obligations into practical regulations, especially as disagreements surface over compliance costs, user experience, and effective safeguards against misleading synthetic media. These discussions highlight the difficulties ahead in crafting a regulatory framework that meets both industry demands and public safety concerns.
As the deadline for the final draft nears, the outcomes of these conversations will significantly impact how AI-generated content is treated across the European Union. Stakeholders must find a way to balance building innovation in AI technologies with making sure consumer protection against potential deception.
The ongoing discussions within these working groups mark a key moment in the evolution of AI regulation, emphasizing the need for continuous collaboration among various sectors to develop solutions that are both practical and effective.
Quick answers
What is the purpose of the Code of Practice on AI-generated content?
The code aims to clarify the obligations for marking, labelling, and disclosing AI-generated content to enhance transparency and consumer protection.
When is the final draft of the code expected to be completed?
The final draft is expected to be completed by early June.
What are the main concerns raised by stakeholders during the workshops?
Industry participants expressed concerns about compliance burdens and feasibility, while civil society called for stronger public interest safeguards.
How will the code affect the use of synthetic media and deepfakes?
The code will establish clear guidelines for marking and disclosing synthetic media and deepfakes, which aims to inform consumers and enhance accountability.
