EU Consultation on Rules for General Purpose AIs Under AI Act

The European Union has launched a consultation on regulations that will be applicable to providers of general purpose AI models (GPAIs) — such as Anthropic, Google, Microsoft, and OpenAI — under the EU AI Act, a risk-based framework for regulating the use of artificial intelligence. Lawmakers aim for the Code of Practice to ensure the development of 'trustworthy' GPAIs by offering guidance to developers on compliance with legal requirements.

The EU AI Act was approved earlier this year and will become effective on August 1. However, implementation deadlines are phased, with Codes of Practice set to come into effect nine months later in April 2025. This allows time for the EU to develop the necessary guidelines.

The European Commission is seeking feedback on the consultation from GPAI providers operating within the EU, as well as from businesses, civil society representatives, rights holders, and academic experts.

The consultation, in the form of a questionnaire, is divided into three sections covering transparency and copyright-related provisions, risk assessment and mitigation for GPAIs with systemic risk, and the reviewing and monitoring of GPAI Codes of Practice.

An initial draft Code will be developed based on the responses and answers provided to the questionnaire.

Individuals and organizations participating in the consultation have the opportunity to influence the development of the template that the AI Office will provide for GPAI providers to comply with the legal obligation to disclose a summary of their model training content.

Further details on the consultation can be found here, with a submission deadline of September 10, 2024, at 6 p.m. CET.

The EU is also seeking expressions of interest to participate in the drafting of the Code through virtual working groups. An iterative drafting process will be used to create the guidance.

The AI Office has invited eligible GPAI model providers, downstream providers, industry organizations, civil society organizations, rights holders, academia, and independent experts to express their interest in contributing to the Code of Practice by August 25, 2024.

GPAI providers will also have the opportunity to attend workshops with meeting chairs and vice-chairs to provide input for each drafting round, in addition to participation in plenary meetings.

The AI Office will appoint meeting chairs and vice-chairs, and it is currently accepting applications from interested independent experts for these roles.

The call for participation and the consultation on the Code follow concerns about the possible exclusion of civil society organizations in the drafting process. The Commission is emphasizing that all interested stakeholders are encouraged to participate, including academia, industry representatives, civil society organizations, rights holders, and public authorities.

EU’s AI Act gets published in bloc’s Official Journal, starting clock on legal deadlines