🤖 EU AI Regulation Update: Proposed Compromise on Foundation Models

← Alle Beiträge
Künstliche Intelligenz · EN

🤖 EU AI Regulation Update: Proposed Compromise on Foundation Models

Following recent challenges in the AI Act Trilogue negotiations, the EU Commission has shared a compromise proposal on Sunday addressing disagreements on the regulation of Foundation Models.

🔍 Key Differentiations in the Proposal:

- General-Purpose AI Model: Defined as an AI model capable of diverse tasks, even with self-supervised scale training.
- General-Purpose AI System: An AI system based on a model serving various purposes, for direct use or integration into other AI systems.

🚀 Advancements Beyond: “General-Purpose AI Model with Systemic Risks”:
This subcategory addresses models with capabilities beyond most advanced general-purpose AI models, with criteria including computational power.

📝 Requirements for AI Model Providers:

- Model Cards: Providers must furnish model cards with documentation on the training process and relevant details for downstream development.
- Copyright Compliance: Adherence to the Copyright Directive, including a sufficiently detailed summary of training content.
- Additional for Systemic Risks: Monitoring and information exchange with the Commission and national authorities.

🤝 Codes of Practice:

The proposal suggests the creation of “Codes of Practice” by the Commission in collaboration with stakeholders. These codes aim to update model card contents and help identify systemic risks.

🌐 Impact and Next Steps:

The compromise integrates demands from member states and the EU Parliament, addressing concerns through elements like model cards and Codes of Practice. The acceptance of this compromise in the ongoing Trilogue is yet to be confirmed.

Let’s engage in this vital conversation shaping the future of AI regulation in the EU.

#AIRegulation #AIGovernance

Foto by Christian Lue on Unsplash

Zu diesem Beitrag gibt es auf LinkedIn ein Bild. Auf LinkedIn ansehen →

Auf LinkedIn diskutieren