After Harm: A Plea for Moral Repair after Algorithms Have Failed.

Area of Science:

• AI Ethics and Governance
• Societal Impacts of AI
• Algorithmic Decision-Making

Background:

• Growing societal concerns regarding AI and algorithmic decision-making.
• Current regulatory and scholarly efforts prioritize risk identification and preventative safeguards.
• A focus on prevention overlooks post-harm scenarios and their long-term consequences.

Purpose of the Study:

• To highlight the inattention to post-harm scenarios in AI ethics and governance.
• To introduce the concept of 'algorithmic imprint' for understanding algorithmic harm.
• To explore the necessity of moral repair beyond system decommissioning or decision reversal.

Main Methods:

• Conceptual analysis of AI ethics and governance frameworks.
• Introduction and definition of the 'algorithmic imprint' concept.
• Argumentation for the inadequacy of current responses to algorithmic harm.

Main Results:

• Algorithmic harm is inadequately addressed by current preventative measures.
• The concept of 'algorithmic imprint' offers a new lens for analyzing harm.
• Decommissioning systems or reversing decisions is insufficient for full redress.

Conclusions:

• A significant blind spot exists in AI ethics regarding post-harm scenarios.
• Addressing algorithmic harm requires acknowledging its lasting effects through 'algorithmic imprint'.
• Moral repair is essential for a comprehensive response to algorithmic harm.