相关实验视频
Updated: Jun 12, 2025

06:31
Force and Position Control in Humans - The Role of Augmented Feedback
Published on: June 19, 2016
7.8K
有帮助,无害,诚实? 人工智能调整和安全的社会技术限制通过强化学习从人类反学习
Adam Dahlgren Lindström1, Leila Methnani1, Lea Krause2
1Department of Computing Science, Umeå University, Umeå, 90187 Sweden.
概括
这项研究批评来自人类反的强化学习 (RLHF),因为它将人工智能,如大型语言模型 (LLM),与人类价值观保持一致. 它发现RLHF在捕捉复杂的伦理方面存在重大局限性,影响AI安全.
科学领域:
- 人工智能的人工智能
- 人工智能伦理学
- 社会技术系统 社会技术系统
背景情况:
- 目前的AI调整方法,包括从人类反中强化学习 (RLHF) 和从AI反中强化学习 (RLAIF),旨在使AI系统与人类价值观保持一致.
- 大型语言模型 (LLM) 是这些对齐技术的主要焦点.
研究的目的:
- 批判性地评估RLHF和RLAIF在将AI与人类价值观和意图保持一致方面的有效性.
- 检查AI调整中的"有用,无害和诚实" (HHH) 原则的局限性.
- 识别人工智能调整研究中被忽视的伦理考虑因素.
主要方法:
- 对RLHF的多学科社会技术批评,理论基础和实际实施.
- 分析HHH原则中固有的紧张关系.
- 在人工智能调整中讨论道德权衡.
主要成果:
- RLHF和RLAIF在捕捉人类道德复杂性方面表现出重大缺陷.
- HHH原理呈现出固有的紧张关系,使人工智能对齐变得复杂.
- 诸如用户友好性与欺骗性,灵活性与可解释性等伦理问题经常被忽视.
结论:
- 对人工智能安全和道德有必要采取更广泛的社会技术方法,整合机构,过程和技术设计.
- 人工智能安全应该被确立为一个社会技术学科,承认规范和政治维度.
相关概念视频
Law of Effect
1.3K
B.F. Skinner, a prominent figure in behavioral psychology, introduced operant conditioning by emphasizing the role of consequences in shaping behavior. This theory builds upon the law of effect proposed by Edward Thorndike, which posits that behaviors followed by satisfying outcomes are likely to be repeated. In contrast, those followed by unsatisfying outcomes are less likely to recur.
Edward Thorndike's foundational work involved studying learning in animals, particularly using puzzle...
Edward Thorndike's foundational work involved studying learning in animals, particularly using puzzle...
1.3K
Stereotype Content Model
14.0K
The Stereotype Content Model (SCM) was first proposed by Susan Fiske and her colleagues (Fiske, Cuddy, Glick & Xu, 2002; see also Fiske, 2012 and Fiske, 2017). The SCM specifies that when someone encounters a new group, they will stereotype them based on two metrics: warmth—or that group’s perceived intent, and how likely they are to provide help or inflict harm—and competence—or their ability to carry out that objective. Depending on the warmth-competence...
14.0K
Current Trends in Nursing II
1.2K
Trends in nursing are multifactorial and associated with changes in society, within the nursing profession, and in other professions. Notably, telehealth and remote nursing contribute to successful healthcare delivery for numerous patients and help reduce stress for nurses due to nursing shortages. Nurses can reach patients, monitor their conditions, and interact with them using computers, audio, visual accessories, and telephones—for example, remote patient monitoring systems. Likewise,...
1.2K

