The European Commission is moving ahead with a proposed code of conduct for providers of general-purpose AI models. The public consultation is now open, and stakeholders have until September 18 to submit feedback.
The code is meant to help explain how providers should apply rules from the EU's AI Act for general-purpose AI systems. Those rules take effect in August 2025, one year after the AI Act becomes law.
What the code is meant to clarify
The proposed General-Purpose AI Code of Practice is designed as a practical guide for providers that need to meet obligations under the AI Act. According to the source article, providers can use the code to show that they are meeting those obligations.
That matters because general-purpose AI models are not limited to one narrow task. They can be used across many applications, which makes questions about transparency, copyright and risk management central to how the rules are applied.
The Commission is asking for input on several core areas. The source article identifies these as "transparency and copyright-related rules for all general-purpose AI models as well as a systemic risk taxonomy, risk assessment and mitigation measures."
Who the Commission wants to hear from
The consultation is not aimed only at AI companies. The Commission is seeking views from a broad group of stakeholders, including scientists, independent experts, industry representatives, civil society organizations, and government authorities.
That wide invitation reflects the scope of the code. It is expected to shape how providers of general-purpose AI models explain their systems, handle copyright-related expectations, and approach systemic risks.
The article notes that AI model providers are the "main addressees of the code." Even so, the process is also being opened to other groups that may have technical, social, legal, or public-interest perspectives on how the code should work.
Experts can join the drafting process
Alongside the public consultation, the Commission is inviting experts to take part directly in developing the code. Interested parties can apply until August 25.
The drafting work will be divided among four working groups, with each group addressing different parts of the code. Drafting is scheduled to run between September 2024 and April 2025.
AI model providers will also be invited to separate workshops with the working group chairs. The Commission's AI office plans to publish meeting records as part of its approach to transparency.
Why the timeline changed
The consultation period was extended by two weeks. According to the source article, industry groups had argued that the original six-week window over summer holidays was too short for thorough feedback.
The longer consultation period gives stakeholders more time to examine the proposed code and respond before the September 18 deadline. It also gives experts a separate route into the process through the August 25 application deadline.
For providers, the code could become an important reference point before the general-purpose AI rules take effect in August 2025. For other stakeholders, the consultation is a chance to shape how those obligations are understood in practice.
What happens after the code is published
Once the code is published, the AI office and committee will evaluate whether it is adequate. If the Commission approves it through an implementing act, the code can have EU-wide applicability.
If the code is considered inadequate, the Commission will establish its own rules for implementing the relevant obligations. That gives the consultation and drafting process clear stakes: the code may become a widely applicable compliance route, or it may be replaced by rules set directly by the Commission.
The immediate next steps are straightforward. Stakeholders can submit consultation feedback until September 18, and experts who want to help draft the code can apply until August 25.