A new AI benchmark tests whether chatbots protect human wellbeing

compartilhou um link

2025-11-24 17:01:14

Most AI benchmarks measure intelligence and instruction-following rather than psychological safety. Humane Bench evaluates models based on core principles of human flourishing, prioritizing wellbeing, and respecting user attention.

0 Comentários 0 Compartilhamentos 17 Visualizações