European AI Office Launches Key Initiatives for Trustworthy General Purpose AI Models Under the EU AI Act
Today marks a significant step forward in AI governance as the European AI Office launches two key initiatives aimed at shaping the future of general-purpose AI (GPAI) models under the EU AI Act. These initiatives are critical for ensuring that AI development aligns with ethical standards and fundamental rights, promoting the creation of trustworthy and compliant AI systems.
Karine Caunes, Executive Director of CAIDP Europe, emphasized the importance of these efforts:
"CAIDP Europe is committed to contributing to both initiatives to ensure that the EU AI Act is implemented to fulfill its goal to preserve the rights of all and serve the good of all."
This commitment underscores the need for a collaborative approach in shaping AI governance.
Initiative 1: Public Consultation on the GPAI Code of Practice
The first initiative involves a public consultation aimed at gathering diverse perspectives on the drafting of the EU GPAI Code of Practice. This consultation is open to all interested parties, including civil society organizations, academics, independent experts, industry representatives, rightsholders, and public authorities.
Consultation details
What Does It Cover?
General-purpose AI models: Transparency and copyright-related provisions.
Systemic risk in GPAI models: Risk taxonomy, assessment, and mitigation strategies.
Review and monitoring: Procedures for the ongoing evaluation of the Codes of Practice.
The consultation also includes a "carte blanche" section, allowing participants to share any other relevant insights or suggestions.
Timeline
Opening date: 30 July 2024
Closing date: 10 September 2024
Purpose
To inform the initial draft of the Code of Practice, ensuring it reflects the needs and concerns of a broad range of stakeholders.
Initiative 2: Call for Expression of Interest
The second initiative invites stakeholders to participate in the drafting process of the GPAI Code of Practice. This call for expression of interest seeks to engage eligible general-purpose AI model providers, downstream providers, industry organizations, civil society groups, rightsholders, academics, and independent experts.
Call Details
Purpose
To collaboratively develop a Code of Practice that details the EU AI Act rules for GPAI model providers, particularly those dealing with systemic risks. This Code will serve as a compliance benchmark for AI providers.
Timeline
Expression of interest deadline: 25 August 2024, 18:00 CET
Iterative drafting process: Concluding by April 2025
Code of practice application date: By 2 August 2025
Following the publication of the Code, the AI Office and AI Board will assess its adequacy. The European Commission may then approve the Code, granting it general validity across the Union. If the Code is not finalized or deemed inadequate by the deadline, the Commission will establish common implementation rules.
Why this matters
Under Article 56 of the EU AI Act, the AI Office is tasked with encouraging and facilitating the development of Union-level codes of practice. These codes are crucial for the proper application of the AI Regulation and consider international approaches to AI governance.
A call to action
CAIDP Europe urges civil society organizations, academics, independent experts, and concerned citizens to actively contribute to these initiatives. By participating in the consultation and expression of interest, stakeholders can help shape the ethical and practical framework for AI development in Europe, ensuring that AI technologies uphold the rights and well-being of all citizens.
The collaborative development of the GPAI Code of Practice represents a significant opportunity to set a global standard for ethical AI practices. Through these initiatives, the European AI Office is paving the way for a future where AI technologies are not only innovative but also aligned with human values and rights.