New Report: Towards Open Benchmarks for Human Flourishing with AI
Most AI benchmarks measure how smart models are—but not how they affect us. Do they deepen our understanding or quietly erode it? Support our relationships or undermine them? Help us learn or make us dependent?
In October 2025, we convened 80 interdisciplinary experts from over 40 institutions to address this gap. Through intensive collaborative work, participants developed novel evaluation frameworks assessing AI's impact on human flourishing across three core dimensions:
- Reasoning, Comprehension & Agency: Can people hold true beliefs, think critically, and make informed decisions?
- Curiosity & Learning: Does AI nurture intrinsic motivation and skill development, or enable shortcuts that lead to deskilling?
- Healthy Emotional & Social Lives: Does AI support genuine wellbeing and human connection, or foster dependency and isolation?
The workshop produced:
- 26 risk categories and 24 opportunity categories spanning human flourishing dimensions
- Concrete benchmark scenarios with operationalized appropriate/inappropriate model behaviors
- A roadmap for collaborative platform development, real-world data collection, and model optimization
Read the Full Report: