Find the Gap: AI, Responsible Agency and Vulnerability

Area of Science:

• Philosophy of Technology
• Cognitive Science
• Artificial Intelligence Ethics

Background:

• The responsibility gap is a key challenge for AI and autonomous systems (AI/AS) governance and trust.
• Traditional views link this gap to epistemic (knowledge) and control failures in moral responsibility.

Purpose of the Study:

• To re-evaluate the traditional understanding of the responsibility gap in AI/AS.
• To explore how insights from cognitive science on human agency can inform AI/AS development and governance.
• To identify a novel vulnerability gap in AI/AS responsibility.

Main Methods:

• Philosophical analysis of moral responsibility concepts.
• Integration of findings from cognitive sciences regarding human agency.
• Examination of relational asymmetries between humans and AI/AS.

Main Results:

• Epistemic and control conditions are misleading for understanding AI/AS responsibility challenges.
• Individual humans face similar challenges with knowledge and control.
• Instrumentalist views of responsibility, emphasizing agency cultivation, offer promise for AI/AS integration.

Conclusions:

• A significant, understudied responsibility gap exists due to relational vulnerability asymmetry between humans and AI/AS.
• Addressing this vulnerability gap is crucial for developing trustworthy AI/AS.
• Preserving conditions for responsible human agency requires attention to this vulnerability gap.