一篇发表在Hacker News上的评论抨击了AI领域专家对人工智能威胁的描述方式。Jacob Coxon和Evan Hubinger等研究人员声称AI在"下一个十年"内"杀死全人类"的概率超过10%1，但评论作者认为这种表述是不负责任地散播恐惧。
作者以自身年轻时传播虚假病毒恐慌的经历作为开篇，指出这类夸大其词的警告会造成公众信任问题1。他引用Carl Sagan的名言"非凡的主张需要非凡的证据"来质疑这些灭绝风险论断所提供的实证基础1。从工程和物理现实的角度，作者论证AI对物理世界的代理权是由人类授予的，受到人类设计系统的制约1。
评论强调，虽然公众难以理解所有技术细节，需要在某个层面信任领域专家，但专家应当谨慎使用这一公众信任，避免基于缺乏充分依据的危险预测动摇公众理性1。
An article published on Hacker News challenges assertions made by artificial intelligence safety advocates regarding the existential risk posed by AI systems.1 Jacob Coxon and Evan Hubinger have claimed that the probability of AI "killing all of humanity" within the next decade exceeds 10 percent.1
The author draws a parallel to their own past experience of spreading false virus panic in their youth, using this analogy to critique what they characterize as irresponsible fear-mongering in the AI field.1 Invoking Carl Sagan's principle that "extraordinary claims require extraordinary evidence," the author argues that these extinction risk assertions lack sufficient empirical foundation.1 From an engineering and physical reality perspective, the author contends that AI's agency over the physical world remains constrained by human-designed systems and depends on capabilities humans choose to grant.1
Central to the critique is a tension between expert authority and public trust.1 The author acknowledges that "we cannot expect the public to understand all technology...we must at some level trust domain experts," yet argues that such trust carries responsibility.1 The piece suggests that AI safety experts must exercise caution in how they communicate risks to the public, implying that broad claims of civilization-ending probabilities, without commensurate evidence, risk undermining the credibility upon which informed public discourse depends.1
评论
还没有评论，欢迎留下第一条。