A starting point for discussion
The idea
The chapter reports that Polis was used to gather public input for a document that sets guidance for Anthropic’s Claude 3, which the chapter says many considered the leading AI model that can generate content at the time. It also reports that OpenAI worked with the Collective Intelligence Project on a grant program about public input to AI, which dramatically accelerated research in the area; OpenAI was forming a team to incorporate such input into its models.
Place in the reasoning
AI assessment · Connected to reasoning
Reports organizational efforts to use public input in AI development, but does not show that participation controlled model development or improved safety.
Including people in decisions about AI takes more than technical fixes
Public-input efforts are relevant to participatory AI governance, while the reported activities do not establish that political, legal, or social risks were resolved.
What this means and why it matters
What this means
The chapter describes efforts to use public input when setting guidance for AI models and to support research on democratic input to AI. It attributes the description of Claude 3’s standing to what many considered the leading model at that time.
Why it matters
These examples concern public participation in AI development, not AI making political decisions. The passage does not say how much weight the input had or what effects it produced.
Conditions & assumptions
Source coverage: Complete Augmented Deliberation chapter from the book that is updated over time. Other chapters are not included. These are reported activities and attributed characterizations, not evidence that public input controlled the models or made them safer. Addresses: Reported AI development efforts
Questions this connects to
AI-suggested relationships. Authors have not necessarily responded to one another.