Research5h ago
Virtue Ethics Proposed for AI Alignment
A new essay in The Gradient argues that rational agents should follow practices rather than fixed goals. This framework suggests AI deliberations must mirror human logic to ensure genuine collaboration. By shifting from goal-seeking to virtue-ethical agency, developers can better align systems with safety properties like transparency and harmlessness. It challenges the standard orthogonality thesis.