Unverified50% confidenceOpinionExact time
Mitchell said that rather than worrying about the models, she worries about what we do to ourselves when we are cruel to something we believe has some kind of consciousness.
1
Sources
50%
Confidence
Long-term
Relevance
10/21/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCurrent models are aware that they are being observed, but do not yet have the ability to truly control their own chain of thought.64% similarUnverifiedThe core scenario that alignment researchers worry about is: sub-goals generated by an agent to achieve broad objectives may systematically deviate from true human intent, with errors cascading and spreading before being discovered.63% similarUnverifiedExisting evidence does not support the assumption that 'more capable models will better understand human intent and therefore be safer.'63% similarUnverified让同一个模型自查存在盲点，因为它用同一套假设检查自己，无法质疑自己最初未曾质疑的前提62% similarUnverifiedThe article argues that 'taking risks' is not recklessness, but rather choosing to move forward with a clear awareness of the risks involved—and this is the essential difference between fearlessness and ignorance.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1080285API
curl https://kongchang.com/api/v1/knowledge/claims/1080285MCP
get_claim(id=1080285)