Website adheres to final Code of Practice for General-Purpose AI, implying compliance with industry standards for ethical AI development and use.
## EU's General-Purpose AI Code of Practice: A Look at Its Publication, Criticisms, and Implications
The European Union's General-Purpose AI (GPAI) Code of Practice, a voluntary guide to help providers comply with the EU AI Act, was published by the European Commission on July 10, 2025 [1][2]. The Code, which becomes effective on August 2, 2025, focuses on transparency, copyright, and safety and security [1][3]. Although adhering to the Code does not guarantee legal conformity, it aims to provide clarity on compliance requirements.
### Key Features and Controversies
1. **Voluntary Adherence**: The voluntary nature of the Code has raised concerns about robust compliance, as adherence is optional and does not provide a legal presumption of conformity with the EU AI Act [1][2]. This could lead to inconsistent application across various organisations.
2. **Regulated Entities' Influence**: Although the drafting process involved numerous stakeholders, including representatives from EU Member States and civil society, there are concerns about the influence of regulated entities (e.g., technology businesses) on the Code's content [3]. These entities have advocated for clearer guidelines and less burdensome regulations, which may impact the Code's effectiveness in ensuring fair and transparent practices.
3. **Transparency**: The transparency chapter aims to ensure that GPAI model providers are clear about their models' capabilities and limitations. However, since the Code is voluntary, there might be variability in how transparently these models are presented, potentially leaving room for interpretation and inconsistent practices [1].
4. **Timing and Delay**: Originally scheduled for release in May 2025, the Code was delayed until July. This delay has sparked questions about the readiness of organisations to implement the EU AI Act's rules effectively [3].
5. **Safety and Security Measures**: The Code includes mechanisms for public transparency and external assessment, as well as a risk taxonomy that requires monitoring of both specified risks and the identification of other risks [1]. Important safety measures have survived intact in the final version of the Code.
### Moving Forward
As the EU prepares for a year of rules implementation, the Commission's support for the Code is expected to play a crucial role in ensuring its provisions remain effective. To address concerns about the voluntary nature of the Code and potential industry influence, the update process should be designed to be robust against undue influence. The AI Office is anticipated to set out plans for provider adherence to the existing Code and establish a timetable for the update process.
Policymakers will need to be more ambitious in future iterations of the Code to address the emerging impacts of GPAI technologies and ensure that the Code remains a robust tool for describing effective risk mitigation approaches.
[1] European Commission (2025). EU General-Purpose AI Code of Practice. Retrieved from
- The voluntary nature of the EU General-Purpose AI Code of Practice, which was published by the European Commission on July 10, 2025, has raised concerns about robust compliance, especially in light of the potential influence of technology businesses on its content during the drafting process.
- Policymakers will need to be more ambitious in future iterations of the Code to address the emerging impacts of GPAI technologies and ensure that policy-and-legislation, such as the EU AI Act, are effectively implemented, while avoiding controversies like voluntary adherence and industry influence.
