The AI entrepreneurial actor: Towards a theoretical framework for artificial intelligence as entrepreneurial actor

Classical theory assumes upon a human entrepreneur who bears uncertainty, exercises alertness and recognises opportunity. We examine what becomes of that assumption when artificial intelligence (AI) performs these issues. Defining AI as agentic systems built on large foundation models, we analyse how far such systems manifest core entrepreneurial attributes and where they reach fundamental limits. We propose the AI Entrepreneurial Agency Framework (AEAF), a four-level model organised along two conceptually independent dimensions: the autonomy of AI judgment and the degree of Knightian uncertainty in the task environment. Three falsifiable propositions theorise the redistribution of uncertainty and accountability, the intentionality boundary on opportunity framing and alertness, and the cognitive-computational re-conception of the individual–opportunity nexus. Incorporating Bayesian entrepreneurship, we specify where AI accelerates experimentation and where it cannot. A structured review of fourteen studies positions the AEAF as an integrated platform for this frontier, with implications for theory, strategy and AI governance.