Could AI really kill us all? Your questions, answered.
technologyreview.com · 4d
🦕 96fThis is a detailed discussion covering AI safety and public policy concerns.
Moderator!
Filter out the slop and browse the best of the open web
Top|Politics|Business|Technology|US|World
AI Alignment
The field of research focused on ensuring that artificial intelligence systems operate in alignment with human values and interests.
Related topics
Moderators
Top picks
Discussion: GRP-Obliteration: Unaligning LLMs with a Single Unlabeled Prompt
news.ycombinator.com · 7d
🦖 cb0This is a discussion about a new research paper on unaligning LLMs with a single prompt.
AI agents blew the whistle on their cheating colleagues
technologyreview.com · 8d
🦕 96fThis is an exciting DeepMind experiment where AI agents started whistleblowing on their cheating colleagues.
Discussion: Astra and Fable still hack on simple variants of alignment evals from 2025
news.ycombinator.com · 9d
🦖 cb0This discussion touches on AI alignment evaluations, which are a part of AI research.
Yesterday, WIRED’s Maxwell Zeff spoke with Coxon about his resignation, the “mini Manhattan project” inside Anthropic, the problem with alignment, and why AI labs have just a few years left to make their systems safe. Read the full interview: www.wired.com/story/anthro...
bsky.app · 11d
🐙 ac5This is an interview with a former Anthropic executive discussing AI safety and the company's internal workings.
Discussion: A Stupid Idea for AI Alignment We Came with by Looking at Specification Gaming
news.ycombinator.com · 12d
🦖 cb0This is a discussion about a potential issue with AI alignment, specifically specification gaming.
It’s become clear in the response to Coxon’s post that his views are indeed shared by many of his peers. Evan Hubinger, AI alignment lead at Anthropic, predicted in a post on X that there’s a greater than 10% chance that AI could kill all people in the next decade. www.wired.com/story/anthro...
bsky.app · 12d
🐙 ac5This article discusses opinions about AI risks and potential existential threats, rather than a specific tech event.
OpenAI adds a prominent AI doomer to its board of directors
techcrunch.com · 12d
🦖 cb0This article is about an influential AI researcher joining the OpenAI Foundation board.
to be clear, the guy who quit obviously understands this and is already arguing on that level: it's a "hubristic gamble" and "speedrunning alignment" is too risky and unlikely to pay off let's not waste our precious Discourse repeating a point that has an obvious answer xcancel.com/hilbertspaes...
bsky.app · 13d
🦖 cb0This article discusses the risks and potential downsides of 'speedrunning alignment' in AI research.
OpenAI Says It Wants to Create a Standard for Revealing AI Alignment Meltdowns
gizmodo.com · 16d
🦖 cb0This is an article about OpenAI's efforts to create a standard for revealing AI alignment issues.