EU Finalises AI Act Transparency Code

The European Union has finalised its Code of Practice under Article 50 of the AI Act. This voluntary code addresses transparency requirements for general-purpose AI (GPAI) models capable of generating text, audio, or visual content. The code aims to increase public confidence in AI technologies.

Adherence to this Code of Practice will be reported annually. Signatories must implement technical solutions for detecting AI-generated content. They are also required to disclose when content has been generated or manipulated by AI systems. The Code outlines specific measures for content authentication and identification.

Signatories must also incorporate mechanisms for detecting illicit use of their AI systems. This includes reporting any identified illicit activities to relevant authorities. The Code seeks to balance innovation with responsibility in the rapidly evolving AI landscape.

Initial signatories include providers of GPAI models. Compliance with the Code of Practice is expected to facilitate adherence to future mandatory requirements of the AI Act. The Code will also be reviewed and updated to reflect technological advancements and regulatory developments. Its principles align with broader EU objectives for trustworthy AI.

The Code's focus on transparency and detectability anticipates the full implementation of the AI Act. This voluntary framework sets precedents for how AI-generated content will be managed within the EU. The long-term impact on the development and deployment of GPAI models remains to be seen.