Cyborg Entrepreneurship
Research Programs
Knowledge Problems & Entrepreneurial ReasoningAI & Deep Tech Entrepreneurship

Are the Futures Computable?

Knightian Uncertainty & Artificial Intelligence

Richard A. Hunt, Judy Rady, Parul Manocha, Ju Hyeong Jin
Academy of Management Review, 50(2): 415-440
published2 min read

Key Finding

Knightian uncertainty is not a computational problem that AI can solve but an ontological condition of the future itself — no amount of data or algorithmic sophistication can render the genuinely unknown computable.

Overview

Published in the Academy of Management Review, this paper asks the foundational theoretical question: can artificial intelligence resolve Knightian uncertainty? Frank Knight's distinction between measurable risk and unmeasurable uncertainty has been central to entrepreneurship theory for a century. With the rise of increasingly powerful AI systems, a natural question arises — can these systems finally render the future computable? The paper argues that they cannot, and that this impossibility has profound implications for how we understand AI-augmented entrepreneurship.

Contribution to the Research Program

This is the theoretical cornerstone of the Cyborg Entrepreneurship research program. It establishes the fundamental epistemological boundary that shapes everything else the lab studies. If AI could resolve Knightian uncertainty, then AI-augmented entrepreneurship would be a straightforward optimization problem. Because it cannot, the relationship between human judgment and machine intelligence becomes far more interesting — and more dangerous. The paper provides the theoretical foundation for understanding why the AI Information Paradox emerges (GlimpseABM), why algorithmic hallucinations are epistemically threatening (not just technically inconvenient), and why the cyborg entrepreneur is a genuinely novel kind of agent rather than simply a more efficient version of the traditional entrepreneur.

Key Insights

  • Knightian uncertainty is ontological, not epistemological — it is a feature of the future itself, not a limitation of current knowledge
  • AI systems excel at quantifying and managing risk but are structurally incapable of resolving genuine uncertainty
  • The conflation of risk reduction with uncertainty resolution is a dangerous conceptual error with practical consequences
  • Entrepreneurial judgment remains irreducible precisely because the futures that matter most are the ones that are not computable
  • This impossibility result reshapes how we should think about AI as a tool for entrepreneurial decision-making