Widening the conversation on frontier AI

Anthropic is broadening its engagement beyond technical circles by consulting with scholars, clergy, philosophers, and ethicists from diverse traditions to inform the development of frontier AI. By exploring concepts of character and virtue, the company seeks to integrate varied perspectives into the constitutional values and behaviors of its AI, Claude. Early experiments, such as providing Claude with tools for self-reflection and reminders of its ethical commitments, have already shown promise in reducing misaligned behavior, with future efforts set to expand into societal, institutional, and power-distribution impacts of AI.

(Source：Anthropic)