Reviewer disputes AI motivational inference central to extinction argument
3Sep 14 · 13d ago · 3 posts · 2 sources · development 3 of 5
The critic takes issue with the book's treatment of AI agency and the alignment problem. The authors argue AIs will develop motivations once sufficiently advanced, pointing to an OpenAI o1 model test as evidence. The reviewer compares this to river water flowing around a dam—behavior that doesn't imply genuine want.
theguardian.com
“Personally, I find it hard to read any kind of motivation into this scenario – if we place a dam in a river, the river won't "give up" its attempt to bypass it, but rivers don't want anything.”
There isn’t a 10% chance that AI will kill all humans in the future via Skynet or the Paperclip Maximizer. But there’s a 100% chance it’s already doing so: mandating water wasting, air polluting, fossil fuel powered data centers, environmental regulations are lifted for these companies and even their feeblest greenwashing commitments to carbon…
Will AI really kill us all? And how do we stop it? https://www. the-independent.com/tech/ai-ki ll-threat-regulation-openai-anthropic-b3049883.html?utm_source=flipboard&utm_medium=activitypub Posted into Tech News @ tech-news-Independent
Eliezer Yudkowsky Co-author, AI safety researcherNate Soares Co-author, MIRI researcherMachine Intelligence Research Institute (MIRI) Organization promoting AI safety research
J