Algorithmic Decision Impact

Definition

Algorithmic decision impact describes an exposure pathway where harm reaches victims through AI-driven decisions made about them — not through their direct use of the system. The AI system scores, ranks, filters, classifies, or adjudicates, and the resulting decision affects the victim’s opportunities, access, or treatment.

This pathway is characteristic of institutional AI deployments: hiring algorithms that screen candidates, credit scoring systems that determine loan eligibility, welfare adjudication tools that approve or deny benefits, and content moderation systems that remove or suppress user content.

How This Pathway Operates

Algorithmic decision impact typically follows a sequence:

1. An institution deploys an AI system to make or assist decisions about people
2. The system processes data about the victim (often without their knowledge or meaningful consent)
3. The system’s output influences a decision that affects the victim’s access, treatment, or opportunities
4. The victim experiences the consequence of that decision, often without understanding the AI system’s role

The distinguishing feature is that the victim is the subject of the AI decision, not the user of the AI system. The victim may never see the system’s interface, understand its criteria, or know that AI was involved.

Distinguishing From Other Pathways

• vs. Direct Interaction: Algorithmic decision impact does not require the victim to engage with the system. The victim is decided about, not interacting with, the AI.
• vs. Economic Displacement: Algorithmic decision impact involves specific decisions about specific people. Economic displacement involves structural market changes that affect groups.

Incidents involving this exposure pathway are listed in the sidebar.