Generative AI Use in Entrepreneurship: An Integrative Review and an Empowerment-Entrapment Framework

A team of researchers including Jackson G. Lu has published a comprehensive academic review titled 'Generative AI Use in Entrepreneurship: An Integrative Review and an Empowerment-Entrapment Framework' on arXiv. The paper systematically analyzes how generative AI (GenAI) tools like GPT-4 and Claude impact entrepreneurs across four key stages: opportunity recognition, evaluation, resource assembly, and venture growth. The central contribution is their novel 'Empowerment-Entrapment Framework,' which positions GenAI as a double-edged sword, capable of both significantly aiding and subtly hindering entrepreneurial success.

At each stage, the framework identifies specific trade-offs. For instance, during ideation, GenAI can improve venture idea quality by up to 40% through rapid prototyping and market analysis, but it also introduces risks like AI hallucinations and training data biases that can derail concepts. During execution, AI can boost productivity and functional breadth, automating tasks from coding to marketing, but it may also fuel 'workslop'—superficial, low-quality output—and erode the entrepreneur's own critical thinking, learning, and relational networks. The researchers also identify key boundary conditions, such as an entrepreneur's metacognition and domain expertise, that determine whether the empowering or entrapping effects dominate.

Beyond theory, the paper offers concrete, practical implications for founders and investors. It provides a structured lens to audit AI use in a startup, helping teams deploy tools like GitHub Copilot or Midjourney more strategically while instituting guardrails against overconfidence and dependency. For the venture ecosystem, it underscores the need for updated training and due diligence processes that account for AI's dual role as both a powerful accelerator and a potential source of systemic risk in new ventures.