AI-mediated plausibility regimes: Entrepreneurial judgment, epistemic risk, and the distribution of entrepreneurial futures

Artificial intelligence (AI) is widely portrayed as enhancing entrepreneurial decision-making through improved prediction, pattern recognition, and data-driven evaluation. Yet entrepreneurship frequently unfolds under conditions of Knightian uncertainty, where action depends on entrepreneurial judgment about uncertain and evolving futures rather than calculable prediction alone. We theorize how AI-based systems reshape the epistemic conditions under which entrepreneurial beliefs are formed and evaluated. We argue that AI systems are not epistemically neutral: by generating plausibility signals from historical data, they influence which ventures appear credible and worthy of support. We conceptualize these dynamics as AI-mediated plausibility regimes and identify a dual epistemic risk. In generative modes, AI may inflate evaluative confidence beyond what the available epistemic grounding warrants (false positives). In evaluative and screening modes, AI may disadvantage low-precedent but potentially transformative ventures (false negatives). By distinguishing these system classes and their distributional effects, we specify when AI may expand entrepreneurial imagination and when it may narrow entrepreneurial variety.