AI Model Welfare — AI Dictionary

The unsettled, disputed question of whether an AI model can have morally relevant states worth protecting, and the practice of labs publishing assessments addressing it.