European Union: Published first draft of General-Purpose AI Code of Practice, including quality of service requirement

On 14 November 2024, the European AI Office published the first draft of the General-Purpose Artificial Intelligence (AI) Code of Practice (the “Code”). It sets forth detailed obligations under the Safety and Security Framework (SSF) for providers of models with systemic risk, requiring continuous assessment, best-in-class evaluation practices, post-deployment monitoring, and mechanisms to ensure performance reliability and risk reduction across the entire lifecycle. The Code mandates documentation on architecture, computational resources, energy consumption, and model limitations, all of which directly impact model reliability and operational integrity. Additionally, the Code encourages the establishment of proportionate mitigations tied to severity tiers, requiring Signatories to articulate and continuously review these in their Safety and Security Reports (SSR), thereby establishing a service quality benchmark aligned with Article 55 of the AI Act.