Governance

Downstream Deployer

An organization that uses or integrates an upstream provider's AI system for their own applications or services.

Definition

Downstream Deployer refers to organizations that use, integrate, or deploy AI systems or models obtained from upstream providers for their own applications, products, or services. They occupy a critical position in the AI value chain, translating general-purpose capabilities into specific use cases.

EU AI Act Context:

  • Deployer definition (Article 3(4)): Natural or legal person using AI system under their authority
  • High-risk deployer obligations (Article 26): Use according to instructions, ensure human oversight, monitor operation
  • Integration responsibility: When integrating GPAI into high-risk systems, deployer may become provider

Downstream Deployer Obligations:

  • Follow upstream provider's instructions for use
  • Ensure intended use aligns with system's designed purpose
  • Implement human oversight measures
  • Monitor for incidents and report as required
  • Conduct fundamental rights impact assessment (for certain high-risk uses)
  • Inform individuals subject to AI decisions

When Deployer Becomes Provider:

  • Substantial modification of high-risk system
  • Changing intended purpose to high-risk application
  • Placing system on market under own name

Related concepts: Upstream Provider, Deployer, Provider, AI Value Chain

Sources

  • EU AI Act Article 26
  • Supply Chain Standards