An intervention to shape policy dialogue, communication, and AI research norms for AI safety
The full article may be found on the original post.
Abstract
Discourse on AI safety suffers from heated disagreement between those sceptical and those concerned about existential risk from AI. Framing discussion using strategic choice of language is a subtle but potentially powerful method to shape the direction of AI policy and AI research communities. It is argued here that the AI safety community is committing the dual error of frequently using language that hinders constructive dialogue and missing the opportunity to frame discussion using language that assists their aims. It is suggested that the community amend usage of the term ‘AI risk’ and employ more widely the ‘AI accidents’ frame in order to improve external communication, AI policy discussion, and AI research norms.