Second Draft of the General-Purpose AI Code of Practice published

The second draft of the General-Purpose AI Code of Practice has been released, reflecting the views of approximately 1000 stakeholders, including representatives from EU Member States and international observers. This draft was informed by feedback and discussions from Working Group meetings and written submissions, aiming to refine the initial draft published on 14 November 2024. The Code serves as a guiding document for AI model providers to demonstrate compliance with the AI Act, focusing on transparency, copyright obligations, and systemic risk management.

The draft was presented to several key bodies, including the European AI Board Steering Group and the European Parliament’s joint IMCO-LIBE Working Group. It outlines obligations for AI model providers, particularly those with models posing systemic risks. The structure of the Code has been revised to clarify objectives, commitments, and measures, and includes preliminary examples of Key Performance Indicators (KPIs). The aim is to balance clear commitments with flexibility, considering the evolving state of AI technology.

Feedback on the second draft is being collected until 15 January 2025, with further discussions scheduled in January. The Working Group meetings will address technical risk mitigation, transparency, risk assessment, governance, and copyright-related rules. Workshops with AI model providers and Member State representatives are planned, with the third draft expected by 17 February 2025.

The Code of Practice remains a work in progress, with ongoing efforts to ensure clarity and alignment with the AI Act. The focus is on providing concrete measures for compliance while considering the size and impact of AI model providers. The regulatory approach aims to foster an ecosystem for AI governance and risk management.

Source: Second Draft of the General-Purpose AI Code of Practice published, written by independent experts