Reaction
A model-versus-data-source framing for AI power debates
A post argues that positions on copyright, automation and model welfare reflect competing views of AI outputs' origins.
TLDR
A post argues that many power-related AI debates can be framed around whether outputs come from the model or from data sources. It places pro-copyright and anti-job-loss/automation positions on the data-source side, and anti-copyright and pro-job-loss/automation positions on the model side. It extends the framing to model welfare, sovereign AI and concentration of power.
Combined views
831
2 Sources, first seen ago
