Value Alignment Is a Pseudo Concept - by Zilan Qian

A translation; humanity is not a singe subject, and alignment is not one-way.

As with the previous case, I am translating this article because I think it should not only live within the Chinese internet. I don’t work in this field, so I could not evaluate how much of the criticism presented here is fair. Intuitively, I do disagree with some points here (noted in the footnote). However, I strongly agree with the arguments that humanity is not a single subject and alignment is not a one-way practice. The article below is translated by AI. Footnote annotations and highlights are mine. By Wang Huanchao, Senior Researcher, Tencent Research Institute In May 2024, Ilya…

saved by

• Lailah Nabegu
• Tina

related reading

• The Artificiality of Alignmentjoinreboot.org
• ALIGNMENT - by vincent huang - a slice of my mindmindslice.substack.com
• Alignment remains a hard, unsolved problem — LessWronglesswrong.com
• AGI Ruin: A List of Lethalities — LessWronglesswrong.com
• What Does It Mean to Align AI With Human Values? | Quanta Magazinequantamagazine.org
• OpenAI Launches Superalignment Taskforcethezvi.substack.com
• Alignment & Succession: The Two Bars of Alignmentnosetgauge.com
• The case against AI alignment — LessWronglesswrong.com
• Full-Stack Alignment and Thick Models of Valuefull-stack-alignment.ai
• Center for the Alignment of AI Alignment Centersalignmentalignment.ai
• The Problem With the Word ‘Alignment’ — AI Alignment Forumalignmentforum.org
• Sequent: Scale and Automation for Higher Confidence in Alignment — Sequentsequent.org