EU Finalizes General-Purpose AI Code of Practice for Enterprises
The EU's finalized General-Purpose AI Code of Practice provides guidelines for enterprises ahead of the upcoming AI Act regulations.
Key Points
- • The EU has published the General-Purpose AI Code of Practice for enterprises.
- • Compliance with the code is voluntary but recommended for meeting legal obligations.
- • The code includes a model documentation form and sections on copyright and safety.
- • A group of 13 independent experts developed the code with input from various stakeholders.
The European Commission has officially released the General-Purpose AI Code of Practice aimed at assisting businesses in adhering to the forthcoming EU AI Act regulations, which will take effect on August 2, 2025. This comprehensive code addresses critical areas such as transparency, copyright, safety, and security concerning the deployment of AI technologies in Europe.
While compliance with the code is voluntary, the Commission advocates for its adoption as a resource for enterprises to ensure they fulfill their legal obligations. The code incorporates a model documentation form, making up one-third of its nine-page transparency section, which necessitates organizations to detail aspects of their AI models, including data sources, training processes, energy consumption, and acceptable use practices.
In another significant section, the code emphasizes copyright considerations, urging enterprises to respect web protection measures to prevent the unauthorized use of copyrighted materials. The most substantial part of the code dedicates 40 pages to safety and security, outlining necessary measures to mitigate systemic risks associated with AI models and detailing responsibilities and incident reporting protocols.
Developed by a panel of 13 independent experts in collaboration with various stakeholders, including model providers and academics, the code represents a substantial step toward establishing a structured framework for AI governance. The broader industry has responded positively, despite some earlier calls for a postponement of the AI Act's implementation, which the Commission confirmed will proceed as planned, marking a crucial milestone for AI regulations in Europe.
"This code will greatly aid enterprises in navigating the complex landscape of AI ethics and obligations that are forthcoming," stated a Commission spokesperson. The code supports efforts for organizations to maintain essential documentation for a minimum of ten years following the market release of their AI products.