SDP & DAS Joint Webinar
Featuring David Townsend, PhD - Digges Family Professor, Pamplin College, Virginia Tech
Moderator: Manel Baucells - Associate Professor, Darden School of Business
Abstract:
Strategic decision making in technology markets now unfolds at the intersection of human creativity and increasingly autonomous artificial-intelligence (AI) agents. In recent years, as we work to adapt decision strategies to this emerging era of AI, we increasingly confront a landscape in which demand can surge or evaporate overnight, production costs fall along AI-driven learning curves, and algorithmic rivals copy, undercut, or out-innovate incumbents at machine speed. My research work addresses these problems through an in-depth analysis of the implications of problems of Knightian uncertainty for this emerging AI era.
Knightian uncertainty captures the limits of prediction when data, interpretive patterns, and causal links are incomplete and the future is unknowable. In recent research, I revisit Knight’s original work, incorporating insights from his published and unpublished papers, to break Knightian uncertainty down into four critical decision problems:
Actor ignorance arises because even the most data-hungry founders—and their AI copilots—possess only partial, often biased, mental models of demand, cost and the strategic problems they are facing.
Practical indeterminism channels the tiny random shocks—an unexpected GPU shortage, unanticipated tariffs—into a cascading set of obstacles that through AI-optimized supply chains and pricing bots to produce outcomes no probabilistic average could anticipate.
Agentic novelty captures the constant appearance of new players and tactics: generative-AI breakthroughs slash development times, self-tuning algorithms discover unorthodox product features, and software agents negotiate partnerships in milliseconds, altering the very choice set open to human strategists.
Competitive recursion amplifies the problems, generating competitive responses to every move—from an algorithmic price cut to an automated feature rollout—changes the landscape that the next move will be computed on, triggering accelerating moves and countermoves among AI-enabled entrepreneurs.
By examining these four problems in the context of future AI-accelerated competition, my work is focused on exploring how various combinations of human judgement and adaptive algorithms remain robust when ignorance, indeterminism, novelty, and recursion interact. These insights offer decision professionals a richer language for framing decision problems through the lens of a Knightian uncertainty and a toolkit for designing decision strategies that remain resilient when the AI-accelerated future is fundamentally open-ended and unknowable.