“Is Pausing AI Possible?” by Richard Annilo

This post is an attempt to analyze the feasibility of establishing an international pause on frontier AI development. More specifically, it highlights a few reasons why someone might consider pausing to be a viable component of AI safety strategy. Summary I present six reasons for why pausing frontier AI development globally might be a reasonable strategy. 1) We might not have enough time to solve alignment. First, if we consider the problem to be the misalignment of human-level general AI systems, there just might not be another way, because we might not have enough time to solve alignment. Timelines could be less than five years. Leaders of labs are talking about timelines of less than five years. AI experts also don’t rule out short timelines, and Metaculus is even more pessimistic. Work left to solve alignment could be much longer. Likewise, there seems to be much uncertainty around how much [...] ---Outline:(00:19) Summary(03:31) 1. We are running out of time to solve alignment(04:52) How much time do we have?(06:22) How much work is left?(08:20) What risk can we tolerate?(10:35) 2. The U.S. public supports a pause(12:49) 3. We’ve done it before (kinda..)(12:53) Trigger events(16:37) Comparing industry sizes(18:45) 4. There are many ways to pause(21:11) A) A ‘CERN for AI’(22:16) B). Conditional pausing on evals(22:53) C) Banning ‘civilian AI’(23:32) About compute thresholds(25:29) 5. China is aware of safety concerns(29:27) 6. We can convince others later(31:04) What can you do?(33:19) Conclusion(34:09) AcknowledgementsThe original text contained 5 footnotes which were omitted from this narration. --- First published: October 9th, 2024 Source: https://forum.effectivealtruism.org/posts/fKMPa7cxSnBCymuRm/is-pausing-ai-possible --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.