苹果与橘子：人工智能的不可比拟性问题
人工智能在不可比较价值之间的挣扎——例如正义与公共安全之间的决策——对自主智能（agentic AI）构成了重大挑战。艾伦·麦凯（Allan McCay）最近的一篇文章深入探讨了这一问题，强调了人工智能的优化重点可能导致与人类判断不一致的选择。解决这一“不可比较性问题”对于开发能够有效应对复杂且充满价值判断的决策的人工智能系统至关重要。详细内容请见： https://link.springer.com/article/10.1007/s00146-025-02317-7
查看原文
AI's struggle with incommensurable values—decisions like justice versus public safety—poses a significant challenge to agentic AI. Allan McCay's recent article delves into this issue, highlighting how AI's optimization focus may lead to choices misaligned with human judgment. Addressing this 'incommensurability problem' is crucial for developing AI systems that can navigate complex, value-laden decisions effectively. See here: https://link.springer.com/article/10.1007/s00146-025-02317-7