LESSWRONG
LW

AI-Assisted Alignment

Edited by habryka, niplav, et al. last updated 20th May 2025

AI-Assisted Alignment is a cluster of alignment plans that involve AI somehow significantly helping with alignment research. This can include weak tool AI, or more advanced AGI doing original research.

There has been some debate about how practical this alignment approach is.

AI systems will likely try to solve alignment for their modifications and/or successors during a phase of self-improvement.

Other search terms for this tag: AI aligning AI, automated AI alignment, automated alignment research

1

1

Posts tagged AI-Assisted Alignment

64A "Bitter Lesson" Approach to Aligning AGI and ASI

1y

41

41Requirements for a Basin of Attraction to Alignment

2y

12

30The Best Way to Align an LLM: Is Inner Alignment Now a Solved Problem?

3mo

34

131We have to Upgrade

2y

35

98[Link] Why I’m optimistic about OpenAI’s alignment approach

3y

15

14Proposed Alignment Technique: OSNR (Output Sanitization via Noising and Reconstruction) for Safer Usage of Potentially Misaligned AGI

2y

9

NicholasKees, janus

3y

47

107Beliefs and Disagreements about Automating Alignment Research

3y

4

65How to Control an LLM's Behavior (why my P(DOOM) went down)

2y

30

53[Link] A minimal viable product for alignment

3y

38

9Infinite Possibility Space and the Shutdown Problem

3y

0

267Discussion with Nate Soares on a key alignment difficulty

HoldenKarnofsky

2y

43

175[Linkpost] Introducing Superalignment

2y

69

168Davidad's Bold Plan for Alignment: An In-Depth Explanation

Charbel-Raphaël, Gabin

2y

40

160Agentized LLMs will change the alignment landscape

2y

102

Load More (15/106)

Add Posts