Entrepreneurship has traditionally been assumed to require extraordinary creativity. According to the actualization approach, the obstacles to entrepreneurial action do not lie in the scarcity of creative abilities, but in the difficulties of knowing in advance whether novel ideas are opportunities as opposed to figments of the entrepreneurial imagination. We argue that advances in Generative Artificial Intelligence (GenAI) vindicate the democratic vision of the actualization perspective, yet simultaneously transform traditional knowledge problems into a grand epistemological challenge: The abundance of ideas makes it increasingly challenging to know which are more likely to stand for opportunities as opposed to mere figments of artificial imagination. We advance an evolutionary approach to this emergent challenge. An evolutionary focus on variation and selection mechanisms helps transform the twin problems at the core of the epistemological riddle into solutions. First, we explain that the proliferation of ideas improves the chances that more quality opportunities will be artificially detected. Second, we explain that the challenge of telling opportunities from nonopportunities becomes more tractable through a process targeted at the elimination of nonopportunities. Importantly, we argue that, whereas AI systems should lead the variation stage, entrepreneurs should lead the selection stage. The reason is that, whereas machine intelligence excels in creativity, human intelligence is more adept at providing the realistic understanding necessary for grounding artificial imagination within real-world constraints. Our theorization turns the traditional view of entrepreneurship on its head. Advances in GenAI force us to abandon notions of entrepreneurs as extraordinarily gifted discoverers of opportunities, let alone opportunity creators. Instead, they may be more plausibly theorized as the restrainers of artificial creativity, prior to attempts to actualize desirable futures.
