backJust days before the European Union Artificial Intelligence Act (AI Act”) went into force on August 1, 2024, the EU AI Office announced a call for expression of interest to assist with drafting the AI Code of Practice mandated for general-purpose AI models. The stated goal of the EU AI Office is to have a final Code of Practice for those general-purpose AI models prepared by April 2025, through the iterative collaborations of four working groups. Individuals interested in participating have until August 25, 2024 to submit their application.
General-Purpose AI Models
Article 3 of the AI Act defines general-purpose AI models as AI models that display significant generality and are capable of competently performing a wide range of distinct tasks, regardless of the way the model is placed on the market, and which can be integrated into a variety of downstream systems or applications.
Article 53 outlines obligations for providers of general-purpose AI models, set to come into effect August 2, 2025, which include:
- Maintaining up-to-date technical documentation of the model, including its training and testing process and the results of its evaluation;
- Maintaining information and documentation about the model and making it available to providers of AI systems who intend to integrate the general-purpose AI models into their AI systems;
- Put in place a policy to comply with EU law on copyright and related rights; and
- Preparing and publicly making available a summary about the content used for training of the general-purpose AI model.
Providers of general-purpose AI models may demonstrate compliance with these obligations by relying on codes of practice, addressed in Article 56 of the AI Act. It is this yet-to-be-drafted code of practice that the EU AI Office now asks for participants to help draft through a series of plenary sessions over the next nine months.
Code of Practice Plenary Structure
Preparation of Code of Practice envisioned in Article 56 of the AI Act will be overseen and facilitated by the EU AI Office, with the assistance of participants through a Code of Practice Plenary. This Code of Practice Plenary will be structured into four Working Groups to address specific issues raised in the obligations set out in Article 53:
- Working Group 1: Transparency and copyright related rules
- Detailing out documentation to downstream providers and the AI Office on the basis of Annexes XI and XII to the AI Act, policies to be put in place to comply with Union law on copyright and related rights, and making publicly available a summary about the training content.
- Working Group 2: Risk identification and assessment measures
- Detailing the risk taxonomy based on a proposal by the AI Office and identifying and detailing relevant technical risk assessment measures, including model evaluation and adversarial testing.
- Working Group 3: Risk mitigation measures
- Identifying and detailing relevant technical risk mitigation measures, including cybersecurity protection for the general-purpose AI model and the physical infrastructure of the model.
- Working Group 4: Internal risk management and governance for general-purpose AI model providers
- Identifying and detailing policies and procedures to operationalize risk management in internal governance of general-purpose AI model providers, including keeping track of, documenting, and reporting serious incidents and possible corrective measures.
Each Working Group will have a Chair, and as needed, Vice-Chair, appointed by the EU AI Office.
Code of Practice Plenary Timeline
A Kick-Off Plenary session will take place in September 2024 and from there three virtual drafting rounds will take place organized across the four Working Groups. This iterative process, which may include additional dedicated workshops with providers of general-purpose AI models invited as participants, will culminate in a final version of the first Code of Practice presented in a Closing Plenary in April 2025.
At the Closing Plenary providers of general-purpose AI models will have an opportunity to provide feedback, including their likelihood of using the Code of Practice to demonstrate compliance with Article 53.
Once published, the Code of Practice will be assessed by the EU AI Office and AI Board. The EU Commission will then either approve the Code of Practice or provide common rules in its place. Biometrics, if permitted by the European Union or relevant law.
This client alert is part of an ongoing series of client alerts focused on the EU AI Act. King & Spalding will continue to vigilantly monitor developments related to this and other AI-related legislation.
