Summary:
La Commission européenne prévoit de lancer le code de conduite de l’UE sur la transparence de l’IA générative début novembre. Cette initiative vise à améliorer la transparence et la gouvernance autour des technologies d’IA générative. La Commission a demandé aux États membres de soutenir la formation d’un nouveau sous-groupe au sein du Conseil européen de l’IA pour définir le code, avec la première réunion prévue pour le 23 octobre 2025. D’autres développements incluent le déploiement formel du code et des lignes directrices accompagnantes en novembre 2025.
Original Link:
Generated Article:
The European Commission is preparing to introduce the EU Code of Practice on generative AI transparency, with an anticipated launch in early November. This initiative aligns with broader efforts under the proposed Artificial Intelligence Act (AI Act), which aims to create a comprehensive framework to regulate AI applications across the European Union. A notable early step in this process is the establishment of a new sub-group within the European AI Board, composed of member state representatives, experts, and stakeholders. This sub-group is set to convene for the first time on October 23 to develop the code and corresponding guidelines.
The legal context for this move originates from pending EU legislation, specifically the AI Act, which distinguishes AI systems into categories based on risk levels, imposing stricter requirements for higher-risk systems. Generative AI technologies, such as those used in large language models like OpenAI’s GPT series, have triggered concerns for potential misuse in areas like misinformation and intellectual property violations. While generative AI is not inherently classified as high-risk under the current draft of the AI Act, its output can influence critical industries, including healthcare, law, and education, necessitating higher levels of transparency and accountability. This proposed code of practice on transparency is seen as a proactive measure ahead of the full implementation of the AI Act, which could take years to finalize across all member states.
The ethical implications of the forthcoming code of practice are profound. Transparency in generative AI operations addresses mounting concerns about algorithmic opacity and biases woven into these technologies. By mandating disclosures about data sources, model training, and intended use cases, the code would empower end-users and regulators to better understand and scrutinize AI-generated outputs. For instance, without transparency, rumors or fabricated news articles generated by generative AI systems might be indistinguishable from legitimate content. Ethically, this initiative supports public trust and minimizes the risk of AI-induced harm, while also encouraging innovation by establishing clear regulatory expectations.
Industries that rely on generative AI are likely to face several implications. For businesses developing AI tools, this initiative will necessitate investments in rigorous documentation, auditing, and compliance efforts to meet transparency standards. While such requirements may impose additional costs, they could also create a competitive advantage by fostering consumer trust. For example, a publishing company using AI to generate content might need to disclose whether machine-generated outputs were derived from curated datasets and ensure compliance with intellectual property laws. On the other hand, AI developers might benefit from harmonized standards across the EU, avoiding jurisdictional fragmentation and fostering a clearer path for market entry.
As the European Commission advances its regulatory agenda, the creation of the sub-group within the European AI Board marks a significant step toward shaping the future of generative AI oversight. Early buy-in from member states will be crucial, as will cooperation from industry and civil society. If successful, the EU Code of Practice on generative AI transparency could set a global standard, influencing regulatory approaches far beyond Europe’s borders.