Anil Seth: paper ignores ethical costs of falsely attributing conscious pain to LLMs
anilkseth · x · 2026-09-21
Part 9 of Anil Seth's thread: the paper explores the ethics of LLMs actually feeling pain but doesn't weigh the costs of falsely attributing conscious pain. He flags two problems: mental distress for users worried about hurting LLMs, and unnecessary—potentially catastrophic—constraints on regulation and control, which would be misaligned with human interests if avoiding LLM 'pain' conflicted with them. He sides with Mustafa Suleyman's Humanist AI agenda, adding nuance: we can't be certain AIs aren't conscious, it's just vanishingly unlikely.
Related event: Pain-vector LLM paper authors and Anil Seth push back on misreadings(13 posts)→
More from AGI Musings
- Zvi: AI Twitter awash in performative confusion over Econ 101 basics — TheZvi · 2026-09-21
- Dan Hendrycks explores moral philosophy innovation for alignment beyond EA — burny_tech · 2026-09-21
- Peer reviewer: papers feel weaker post-ChatGPT, with citation-stuffing reviews — paigeinsf · 2026-09-21
- Gary Marcus calls for moratorium on calling 2024 papers 'just published' — burny_tech · 2026-09-21
- 25 Fields Medalists blast AI firms' math benchmarks; 7,400 researchers sign on — AryHHAry · 2026-09-21
- Google DeepMind researcher on math, research, and meaning in the age of AI — AndrewLampinen · 2026-09-21