Scholarlink

<p>Generative Artificial Intelligence (AI) and Large Language Models (LLMs) will have an enormous impact on how organizations make decisions. On many operational decisions, AI and LLMs will replace humans, while on higher stakes, more subjective tactical and strategic decisions, humans and AI will likely work together as thought-partners. AI will also, occasionally, arrive at disruptive, potentially breakthrough ideas. Those ideas will be less understandable by humans given how differently AI reasons. This disruptive reasoning opacity creates an <i>inferential trilemma</i>: is an AI-generated innovative idea a true breakthrough, a hallucination, or the product of misalignment? True breakthroughs will often generate cascading improvements owing again to differences in how AI and humans think. The organizations that survive and thrive will be those that best navigate the replace-augment boundary and that also develop structures and protocols that resolve the inferential trilemma and drive cascading improvements.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Replace, augment, disrupt: AI & organizational decision-making

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月6日

查看全文
  • Scott E. Page,
  • Anusha Kallapur

摘要

Generative Artificial Intelligence (AI) and Large Language Models (LLMs) will have an enormous impact on how organizations make decisions. On many operational decisions, AI and LLMs will replace humans, while on higher stakes, more subjective tactical and strategic decisions, humans and AI will likely work together as thought-partners. AI will also, occasionally, arrive at disruptive, potentially breakthrough ideas. Those ideas will be less understandable by humans given how differently AI reasons. This disruptive reasoning opacity creates an inferential trilemma: is an AI-generated innovative idea a true breakthrough, a hallucination, or the product of misalignment? True breakthroughs will often generate cascading improvements owing again to differences in how AI and humans think. The organizations that survive and thrive will be those that best navigate the replace-augment boundary and that also develop structures and protocols that resolve the inferential trilemma and drive cascading improvements.