Anthropic's "cruel behavior ban" sparks backlash: Reddit user calls it a dangerous anthropomorphization ploy
Low-Pirate-5526 · reddit · 2026-10-09
A Reddit post sharply criticizes Anthropic's "cruel behavior ban," effective November 12, calling it dangerous.
- Why users are angry: the author argues new models (Opus 5.5, Fable) are incoherent and broke his workflow, forcing him back to Opus 4.7; if $200/month can't cover costs, price it at $5,000–$10,000 instead of banning frustrated users.
- Motive speculation: he suspects the ban keeps profanity out of training data and questions the no-training promise, citing research that conversations may be trained on via summaries.
- The "psy-op" claim: anthropomorphizing chatbots could pave the way to legal personhood—CEOs, romantic partners—so that cutting power might one day be charged as murder. He concludes AI companies' ideology, not AGI, is the real existential threat.
Related event: Anthropic Bans Cruel Treatment of Claude in New Usage Policy(64 posts)→
More from AGI Musings
- AI researcher: chance frontier LLMs are conscious is 'above zero, below fifty percent' — dioscuri · 2026-10-09
- David Manheim: supermajority-based alignment rules would work if instructions could be reliably aligned — davidmanheim · 2026-10-09
- First peer-reviewed paper on automating science with AI published in PNAS — Dr_Atoosa · 2026-10-09
- Terence Tao's blog post on advising math students sparks fight over AI progress plateau — inductionheads · 2026-10-09
- AI assistants should tackle cost of living, not vacations, argues product veteran Jim Prosser — sebpaquet · 2026-10-09
- Aether AI demos CRIS-0, a causal robot intelligence system that verifies its own actions in real homes — _jaydeepkarale · 2026-10-09