AI Needs Public Feedback At Scale Before It’s Too Late

TL;DR

Historical communication disruptions, like the printing press and mass media, eventually led to new forms of social consensus.
Social media and AI are currently in a chaotic phase, but they are also the most plausible candidates to eventually tame disorder.
The scale of global connectivity and AI presents an opportunity for unprecedented public deliberation.
Tech companies possess the infrastructure to conduct large-scale consultations with millions of users on AI.
AI models currently exhibit biases and may not represent global values; public input is needed for reinforcement learning and accountability.
A multi-tiered approach involving LLM-mediated consultations, online deliberations, and citizens' assemblies can create democratically written global and local constitutions for AI.
Tech leaders at Anthropic, OpenAI, and Meta are urged to utilize their platforms for large-scale public input on AI.

Continuar leyendo el artículo original