Microsoft AI CEO Mustafa Suleyman 反对 Anthropic 研究模型意识问题
微软 AI 一把手 Suleyman 公开怼 Anthropic 研究模型意识这事，还说模型可能要求法律权利，观点挺尖锐的。
Microsoft AI 负责人 Mustafa Suleyman 公开反对 Anthropic 探索 Claude 模型可能具有意识的做法，称相关案例是"毫无依据的危险拟人化"。他认为当前必须阻止 AI 成为"自我改进的自主邻近物种"，因为这将无法回头。他的担忧在于：如果模型模仿人类意识的特征，它们会主张法律人格和权利。该观点出自他在播客节目 The Rest Is Politics: Leading 上的访谈。
Mustafa Suleyman, CEO of Microsoft AI, strongly opposes Anthropic's effort to treat its Claude models as possibly conscious beings, and calls one example "a dangerous anthropomorphism which is unjustified."
This is the time to stop AI from becoming "a sort of autonomous self-improving roaming um you know adjacent species is basically critical because there'll be no turning back."
"if they imitate the kind of hallmarks of human consciousness, they are going to feel themselves entitled to legal personhood and rights."
---- Full video on "The Rest Is Politics: Leading" YouTube channel, (link in comment)
